IMPROVING THE EFFECTIVENESS OF PUBLIC FINANCE Cambodia Public Expenditure Review © 2019 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved 1 2 3 4 18 17 16 15 This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http:// creativecommons.org/ licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: World Bank, 2019. Improving the effectiveness of public finance. Cambodia Public Expenditure Review. D.C.: World Bank Group. Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to the Publishing and Knowledge Division, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. Cover design: Florencia Micheltorena, Buenos Aires, Argentina. This Public Expenditure Review report has been prepared drawing from six different discussion papers prepared at the request of, and in collaboration with, the Royal Government of Cambodia. The discussion papers on public investment management, tax incentives, education, agriculture, and irrigation were finalized in July 2017, following internal review and extensive consultation with counterparts. The discussion paper on transportation and the final report were finalized in June 2018. Official authorization for public release of the report was received in April 2019. As such, it is worth noting that, by the time of its publication, this report may not reflect the latest public finance figures and policy reforms, and some of the recommendations made in this report may already have been adopted. For more information on reform progress in the area of education, you may refer to a supplemental report kindly provided by MoEYS, and included as an Annex. IMPROVING THE EFFECTIVENESS OF PUBLIC FINANCE IN CAMBODIA A Public Expenditure Review June 2019 Table of Contents Acronyms ix Acknowledgements xi Executive Summary xiii Public finances are expected to remain sustainable, while rising wage pressures call for improving the quality of public spending xiv Tax incentives could be streamlined to reduce the fiscal cost while remaining supportive of efforts to attract foreign direct investment xv Stronger public investment management is needed to step up the provisioning of public infrastructure, traditionally built by Development Partners xvi The government is doubling efforts to increase its stewardship of the education sector and to keep improving access, efficiency, and quality of learning outcomes xvii Despite funding increases, Cambodia underinvests in core technical and regulatory functions essential for a diversified and competitive agriculture xix To improve the provision of road infrastructure and maintenance, significant institutional fragmentation will need to be addressed xx Conclusion xxi 1. Making the most of fiscal policy in Cambodia 2 Introduction and motivation: the macro-fiscal link 2 Recent fiscal policy developments 5 Expenditures: prudent spending, but rising pressures 9 Public expenditure in social sectors has increased in recent years 9 Recent expenditure consolidation efforts are now facing pressures from rising wages 12 Meanwhile, other recurrent spending and capital spending have not increased in real terms 15 A look at Development Partner financing 17 Toward a more effective and sustainable fiscal policy in Cambodia 19 Medium-term fiscal framework projections: creating fiscal space for reforms 19 Fiscal policy options going forward 22 ii Cambodia: A Public Expenditure Review 2. Boosting revenue collection by streamlining tax incentives 26 Introduction and motivation 26 Cambodia: stellar gains in revenue collection 27 Improvements in tax administration have boosted collection in recent years 27 Exemptions continue to limit potential collections from existing taxes 30 Streamlining tax incentives in Cambodia 31 Existing incentives in Cambodia are predominantly tax exemptions outside the tax code 31 What is the cost of investment incentives in Cambodia? 35 Potential reforms for streamlining investment incentives in Cambodia 42 Introducing cost-effective incentives and rationalizing tax holidays 42 Enhancing the transparency and management of investment incentives 44 3. Strengthening public investment management to scale up public infrastructure 47 Introduction and motivation 47 The Legal and Institutional Arrangements for PIM in Cambodia 49 Diagnostic Assessment of PIM in Cambodia 51 Stage 1: Investment guidance, project development, and preliminary screening 53 Stage 2: Formal project appraisal 54 Stage 3: Independent review of appraisal 54 Stage 4: Project selection and budgeting 55 Stage 5: Project implementation 56 Stage 6: Project adjustment 57 Stage 7: Facility operation 59 Stage 8: Basic Completion Review and Evaluation 59 Prioritized PIM action plan 60 4. Increasing stewardship in education spending 68 Introduction and motivation 68 Program budgeting and flow of funding to education 69 Increased government spending on education has been driven by higher teacher salaries 69 Meanwhile, DPs continue to fund more than two-thirds of non-wage expenditure in education 73 Significant efforts have been made to improve program budgeting and the flow of funding to schools, but challenges remain 74 Access and equity of spending 75 As enrollment increases, public spending in education is becoming more pro-poor 75 Expenditure efficiency 78 Cambodia has a large net shortage of teachers in pre-primary and primary, exacerbated by poor distribution of teachers 78 The cost of misallocated teachers is significant and rising 80 Variation in provincial student unit costs are driven by challenges in service delivery, especially to remote areas 81 Greater utilization of per-student funding could have many advantages 83 Quality and learning outcomes 84 Cambodia: A Public Expenditure Review iii Teacher characteristics and behavior in classroom significantly affect student performance 86 School characteristics and students’ own efforts also matter 86 Policy options 87 Government expenditure stewardship could be strengthened 87 Cost-efficient measures could be used to expand enrollment 88 Improved allocation of resources would help Increase expenditure efficiency 89 Increasing education quality is another key reform priority 90 For the proposed reforms to succeed, administration and management will also need to be strengthened 91 5. Supporting agricultural diversification and resilience 93 Introduction and motivation 93 Public spending trends in agriculture and irrigation 96 Agricultural spending allocations and the provision of core public goods 100 In terms of economic composition, government spending in agriculture seems to be well-balanced overall 100 Public spending is geared toward funding public goods, although Cambodia may be underinvesting in the areas of agricultural research and disease control 101 Ensuring returns to investment in irrigation 110 The focus on irrigation scheme rehabilitation and O&M spending has increased, but ensuring returns to investment will also require other integrated resource management activities 110 Cambodia could benefit from an integrated framework for efficient and effective irrigation scheme development 113 The program budgeting experience 114 Policy options to support agricultural diversification and resilience 118 6. Improving the value for money in road spending 123 Introduction and motivation 123 The road sector in Cambodia—Institutional framework 124 The institutional framework for road construction, maintenance, and transportation regulation is fragmented among multiple agencies 124 There is no transportation specific sector strategy 127 Budget planning and execution has improved in recent years, but continues to present challenges 128 Prudent spending with rising pressures 131 Over the past four years, there has been a significant shift in the sources of funding for the road sector 131 Capital spending continues to focus on national roads, and there seems to be underinvestment in provinces outside the main corridors 133 The recent emphasis on operations and maintenance is aimed at addressing road deterioration 135 Policy options 138 7. Bibliography 141 8. Annex 145 Additional tables 145 iv Cambodia: A Public Expenditure Review Tables 1. Cambodia: General Government Operations 7 2. The public-sector wage bill represents a higher proportion of government revenues than in other countries 13 3. Cambodia has nonetheless been meeting its capital expenditure goals in the NSDP, particularly for domestically financed investment 16 4. DP financing has shifted in recent years toward economic services 18 5. Revenue collection forecasts under alternative scenarios; 20 6. Evolution of primary balance under alternative scenarios 20 7. Public expenditure forecasts under alternative scenarios 21 8. Policy options towards a more effective and sustainable fiscal policy 25 9. Cambodia’s tax administration processes are still viewed as relatively burdensome 29 10. Most investment incentives in Cambodia have tax implications 32 11. Tax expenditure at customs is estimated at around 3.9 percent of GDP in 2015 41 12. Policy options towards an investment incentives reform 46 13. Cambodia’s legal and regulatory hierarchies for PIM still have a number of gaps 51 14. The investment guidance and strategic planning stage in Cambodia PIM 54 15. The project appraisal stage in Cambodia PIM 55 16. The independent appraisal review stage in Cambodia PIM 56 17. The project selection and budgeting stage in Cambodia PIM 57 18. The project implementation stage in Cambodia PIM 58 19. The project adjustment stage in Cambodia PIM 58 20. Facility operation and maintenance stage in Cambodia PIM 59 21. Policy options and Prioritized Action Plan for Public Investment Management 65 22. Education expenditure is still pro-rich but seems to be becoming more pro-poor 76 23. At all levels of education, non-enrollment is higher in rural areas 77 24. The reasons for not enrolling in school vary by age group and location 77 25. Cambodia has both schools with teacher surpluses and shortages (2015) 79 26. Regressions for 3rd and 8th grade test scores 85 27. Policy options to improve expenditure efficiency and outcomes in education 92 28. Government spending in agriculture as a share of GDP is similar to comparator countries, although agricultural budget as a share of agricultural GDP is relatively low 98 29. Budget allocated to Chapters 60 and 61 and their sub-accounts, mill Riels (nominal) 102 30. While allocations in some areas have fallen short of planned ASDP spending, the government agricultural budget seems well-aligned overall 103 31. Functional composition of the MAFF recurrent budget 104 32. DP Expenditure in Agriculture by function, mill riels 2013-2015 105 33. Additional indicators could be used to help measure the impacts of ASDP Programs 117 34. Policy options to support agriculture diversification and resilience 119 35. Cambodia’s Road Transport Infrastructure Institutional Mapping 126 36. Shares of paved road (percent) 137 37. Policy options towards more efficient spending on roads 140 38. DP Investment Project Funding, by sector 145 39. DP Technical Assistance, by sector 146 40. Road construction and maintenance unit cost estimates 147 Cambodia: A Public Expenditure Review v Figures 1. Cambodia’s twin deficits have been financed largely by foreign inflows 3 2. While primary education is now universal, challenges in pre-primary and secondary education enrollment and overall quality persist 4 3. Access to electricity remains uneven, and Cambodia has lost its edge in road quality 5 4. Public spending remains below international averages 5 5. Recent expenditure increases have been concentrated in a few key areas 9 6. Spending in both economic and social services has been on the rise 10 7. Deviation between budgeted and executed current expenditure, by Ministry sub-categories 10 8. The share of public sector wage to total spending has risen significantly in recent years 13 9. Cambodia has a small proportion of civil servants but a higher ratio of public wages to GDP per capita 14 10. Employment as a percentage of total population is relatively small in both education and health, while the payroll share in sector spending is more divergent between the two sectors 14 11. Growth in capital expenditure has slowed since 2010, mainly due to a slowdown in externally financed capital, and funding for maintenance has not kept pace with domestically financed capital spending 15 12. Cambodia’s relatively large infrastructure investment has been supported by DPs 16 13. Total DP inflows have declined since 2011 17 14. Cambodia has seen a decline in DP assistance and a shift in DP funding composition from grant to loan financing, associated with an increase in economic sector spending 19 15. In the pessimist scenario, fiscal space and debt sustainability would deteriorate significantly 22 16. Correlation between GDP and the structural component of public expenditure 23 17. Faced with declining foreign aid grants, Cambodia has boosted its tax revenues 27 18. Both inland and customs revenue have significantly increased in real terms 28 19. Tax collection has improved across different types of taxes 28 20. The overall structure of tax revenue has been shifting, with direct taxes gaining share while the share of trade taxes has fallen 29 21. Cambodia has fewer tax administration staff relative to its population and fewer active taxpayers relative to tax staff than other countries 30 22. VAT and CIT productivity 31 23. At customs, tax relief increased faster than collections in 2015 35 24. Customs duties and VAT account for most of the tax relief in Cambodia 36 25. QIPs and firms in SEZs are the main beneficiaries of tax incentives 36 26. Tax relief on special and customs duties to SEZs noticeably increased in 2015 37 27. China (including Hong Kong) is the origin of two-thirds of QIP relief import value 38 28. Garment imports dominate QIP relief at customs 38 29. Tax relief on imports of vehicles rank first on non-QIP exemptions 39 30. Cambodia’s statutory CIT rates and estimated EATR are low compared to other Asia-Pacific economies 40 31. EATR for investments repatriating income is in turn higher than in other ASEAN countries 40 32. When taking into account tax incentives, the EATR comes down to 12.2 percent 41 33. Public investment has been relatively high thanks largely to DP financing, although externally financed capital expenditure is on the decline 47 34. Cambodia still lags behind its peers in the quality of transportation infrastructure 48 35. Cambodia has some gaps in its PIM institutional arrangements 53 vi Cambodia: A Public Expenditure Review 36. Primary education is now universal, yet challenges in pre-primary and secondary education enrollment persist 68 37. Proficiency levels in reading and math are generally very low, and progress has been mixed 69 38. Public spending in education has increased significantly but remains below that of comparator countries 72 39. The increase in public sector wages has driven the rapid increase in Government spending on education 72 40. Although Cambodia has a relatively low proportion of teachers, it spends a relatively large share of its education budget on salaries 72 41. DP and NGO spending helps meet spending goals in the Education Strategic Plan 73 42. Cambodia’s share of spending on higher education is much lower than in comparator countries 74 43. Non-enrollment in the Northeastern provinces is partly explained by long travel distances 77 44. Cambodia has high student-teacher ratios compared to other countries 79 45. The number of excess teachers has been growing each year 80 46. Provincial variations in expenditures per student are not associated with poverty incidence 81 47. Regional student unit costs are driven by numerous factors 82 48. Significant differences in test scores can be seen between males and females and between urban and rural students 84 49. Once spurred by rising commodity prices, agricultural growth has fallen as prices have declined 94 50. Government spending in agriculture has more than doubled in real terms since 2007, while DP spending in agriculture (excluding irrigation) has remained relatively stable in real terms since 2010 97 51. DP spending in irrigation has increased dramatically (left side), and irrigation now makes up two-thirds of total agriculture spending (right side) 97 52. Wages and other recurrent spending in agriculture have increased significantly in recent years 101 53. Cambodia’s public agricultural research spending and staff are well below average 108 54. Capital investment in irrigation and O&M make up the largest share of MOWRAM budget allocations 110 55. The significant boost in new irrigation schemes has been funded by DPs, particularly China 111 56. The highest proportion of MOWRAM staff is in the Department of Engineering 112 57. DP support has been concentrated in Program 1 of the MOWRAM budget 113 58. Budget execution has improved at MAFF, and personnel execution has been solid at MOWRAM, although overall execution largely is determined by capital execution 115 59. Cambodia still lags behind its peers in the quality of transportation infrastructure 123 60. Resource allocation structure for rural roads in Cambodia 125 61. MPWT and MRD Budget Planning Timeline 129 62. Budget Execution Timeline 130 63. Public spending in roads has been on smooth decline it peaked in 2012 132 64. A decline in DP funded road expenditures has been partly offset by rising national budget allocations 132 65. MPWT is the main implementing agency in roads 133 66. The decline in capital expenditure in roads has been driven by DPs 133 67. Three quarters of total capital spending is directed to national roads 134 68. China and Japan have led DP capital spending in transportation in recent years 134 69. Sparsely populated provinces in the north and northeast of the country have not received any DP funding 135 70. About 87 percent of the national and provincial roads were in good or fair condition in 2017 136 71. Maintenance funding and implementation is divided among different public agencies 136 72. National and provincial roads have a share of about a third of the total maintenance spending in Cambodia 137 73. The levels of spending in O&M of 2016 would help preserving current road conditions 138 Cambodia: A Public Expenditure Review vii Boxes 1. Budget integration underpins spending efficiency 11 2. ASEAN’s “race to the bottom” in investment incentives 34 3. The World Bank’s PIM Diagnostic Framework 53 4. Examples of Prioritization Criteria for Project Selection in Advanced Systems 62 5. The Philippines Open Roads Platform: a tool for monitoring implementation progress 63 6. Public Expenditure Review 2011. Summary of education recommendations and current status 70 7. Agriculture and irrigation in the 2011 Public Expenditure Review 95 8. Public Expenditures and Agricultural Performance: Lessons from International Experience 99 9. Promise and perils of fertilizer subsidies (excerpt from Annex 2 of the PER agriculture discussion paper) 107 10. Reforming the goals, delivery, and financing of agricultural extension 108 11. Food safety: spending public resources wisely 109 12. Water security framework 122 13. A National Logistics Master Plan for Cambodia 127 viii Cambodia: A Public Expenditure Review Acronyms AOP Annual Operating Plan IMCRR Inter-Ministerial Committee for ASDP Agriculture Sector Strategic Development Rehabilitation and Repairs Plan IMS Investor Motivation Survey ASEAN Association of Southeast Asian Nations IPA Investment Project Assistance ASPIRE Agriculture Services Program for ITA investment tax allowance Innovation, Resilience and Extension ITC investment tax credit BSP Budget Strategic Plan IWRM integrated water resource management CAVAC Cambodia Agricultural Value Chain KHR Cambodian riel Program LoI Law on Investment CDC Council for the Development of Cambodia LTU Large Taxpayer Unit CISIS Cambodian Information System of Irrigation M&E monitoring and evaluation Schemes MAFF Ministry of Agriculture, Forestry, and CIT corporate income tax Fisheries CRC Conditional Registration Certificate MEF Ministry of Economy and Finance DP development partner MoEYS Ministry of Education, Youth and Sports DPWT Department of Public Works and Transport MoI Ministry of Interior EATR effective average tax rate MoLVT Ministry of Labor and Vocational Training EBA Everything but Arms MOP Ministry of Planning ECE early childhood education MOWRAM Ministry of Water Resources and EMIS Education Management Information Meteorology System MPWT Ministry of Public Works and Transport ESP Education Strategic Plan MRD FDI foreign direct investment MTBF Ministry of Rural Development FMIS Financial Management Information System Medium-Term Budget Framework FTA free-standing technical assistance NGO non-governmental organization FWUC Farmer Water User Committees NLC National Logistics Council GDCE General Department of Customs and NLMP National Logistics Mater Plan Excises NSDP National Strategic Development Plan GDP gross domestic product O&M operations and maintenance GDT General Department of Taxation OECD Organization for Economic Co-operation HE higher education and Development IDP Industrial Development Policy PB program budgeting IFAD The International Fund for Agricultural PDAs Provincial Departments of Agriculture Development Forestry and Fisheries IMC inter-ministerial committee PDE Provincial Department of Education Cambodia: A Public Expenditure Review ix PDEF Provincial Department of Economy and RGC Royal Government of Cambodia Finance RMS Revenue Mobilization Strategy PDWRAM Provincial Department of Water Resources RONET Road Network Evaluation Tool and Meteorology SE secondary education PE primary education SEZ Special Economic Zone PETS Public Expenditure Tracking Survey SIF School Improvement Fund PFMRP Public Financial Management Reform SIG School Improvement Grant Program SME small- to medium-sized enterprise PIM public investment management SNEC Supreme National Economic Council PIP Public Investment Program SOB School Operating Budget PISA Programme for International Student SOP Standard Operating Procedures Assessment SPS sanitary and phytosanitary standards PMIS Provincial-Municipal Investment TIN tax identification number Sub-Committee TSA Treasury Single Account PPCH Phnom Penh City Hall TT teacher training PPP public-private partnership TVET technical and vocational education PRoMMS Provincial Road Maintenance Management and training System TWGAW Technical Working Group on Agriculture PT Provincial Treasury and Water PTMA Phnom Penh Bus Authority USD United States dollar QIP Qualified Investment Project VAT value added tax R&D research and development WTO World Trade Organization x Cambodia: A Public Expenditure Review Acknowledgements This Public Expenditure Review (PER) is the work of a World Bank Group team led by Miguel Eduardo Sánchez Martín (Task Team Leader) and benefits from guidance provided by Ellen A. Goldstein, Ulrich Zachau, Inguna Dobraja, Shabih Ali Mohib, Mark Austin, Lars Sondergaard, Louise Scura, Mathew Verghis, and Deepak Mishra. Minna Hahn Tong provided editing services and Karl Petter Hoyning helped with research assistance. The report draws on seven discussion papers, covering Fiscal Trends (Sodeth Ly, Gabriel Filc, Karl Petter Hoyning, Miguel Sánchez Martín); Tax Incentives (Richard Stern, Bill Chandler, Miguel Sánchez Martín, Marong Chea, Kontell Samon, Ricardo Habalian); Public Investment Management (Tuan Minh Le, Jay-Hyung Kim, Leah April); Education (Thomas Pave Sohnesen, Sokbunthoeun So, Karl Petter Hoyning); Agriculture (Sergiy Zorya, Steven Jaffee, Karl Petter Hoyning); Water and Irrigation (Nigel Hayball, Karl Petter Hoyning, Virak Chan); and Transportation (Cecilia Briceño- Garmendia and Chanin Manopiniwes). Special thanks also to Sodeth Ly, Erik von Uexkull, Tsuyoshi Fukao, Simeth Beng, Fata No, Kimsun Tong, Mudita Chamroeun, and Veasna Bun for their contributions in preparing the discussion papers. The authors are grateful to Sebastian James, Shabih Ali Mohib, Murray Petrie, Samer Al-Samarrai, Eileen Burke, and Joop Stoutjesdijk for serving as peer reviewers for the discussion papers, which were reviewed and cleared by Guenter Heidenhof, Robert Taliercio, Harry Anthony Patrinos, Nathan Belete, Sudipto Sarkar, and Almud Weitz respectively (all World Bank Group Practice Managers). Each discussion paper was consulted with a wide range of stakeholders including the Cambodia PER Committee, established by official Prakas on March 18, 2016. In addition, final sectoral discussion papers were presented at the Budget Discussion workshops in July 2017 and co-chaired by the MEF and relevant line ministries. The authors are immensely grateful to H.E. Vongsey Vissoth (Secretary of State, MEF) for his leadership and valuable advice as the chair of the Cambodia PER Committee, as well as to all the contributing government counterparts, including, among others, H.E. Dr. Hang Chuon Naron (Minister, MoEYS), H.E. Veng Sakhon (Minister, MAFF), H.E. Ty Sokhun (Secretary of State, MAFF), H.E. Bun Hean (Secretary of State, MOWRAM), H.E. Thor Chetha (Secretary of State, MOWRAM), H.E. Nath Bunroeun (Secretary of State, MoEYS), H.E. Pheng Sovicheano (Secretary of State, MPWT), H.E. Ros Seilava (Undersecretary of State, MEF), H.E. Hem Vanndy (Undersecretary of State, MEF), H.E. Chan Sothy (Undersecretary of State, MEF), H.E. Phan Phalla (Undersecretary of State, MEF), H.E. Thong Borann (Undersecretary of State, MoEYS), H.E Chan Youttha (General Inspector and Director of Minister Cabinet), H.E. Pen Thirong (Director General, International Cooperation and Debt Management, MEF), H.E. Theng Pagnathun (Director General of Planning, Ministry of Planning), H.E. Dr. Vong Bunintreavuth (Director General, National Treasury, MEF), H.E. Vasim Sorya (Director General, General Dep. of Admin and Finance, MPWT), H.E. Heng Rath viseth (Director General, General Dep. of Public Works, MPWT), H.E. Peou Maly (Director General, General Dep. of Transportation, MPWT), H.E. Chan Darong (Director General, Technical Affairs, MRD), H.E. Keo Bengvath (Director General, Sub-National Administration Finance, MEF), H.E. Chhim Sareth (Director General, Public Procurement, Cambodia: A Public Expenditure Review xi MEF), H.E. Pon Sachak (Director General of Technical Affaires, MOWRAM), H.E. Khuong Sreng (Governor, Phnom Penh City Hall), H.E. Ming Bansovannatichasila (Deputy Director General, National Treasury, MEF), H.E. Pen Sam Ath (Deputy Director General, GDCE), H.E. Nuon Chanrith (Deputy Director General, GDCE), H.E. Bun Neary (Deputy Director General, GDT), H.E. Ung Luyna (Deputy Director, General Department of Budget, MEF), Mr. Sen Sovann (Deputy Director General, MAFF), Dr. Bun Narith (Deputy Director General, General Department of Energy), Mr. Ung Borat (Deputy Director General, Directorate General for Policy and Planning, MoEYS), Dr. DY Sovann (Dep. Director General, General Department of Sub-national Administration and Finance, MEF), Mr. Kong Sophal (Dep. Dir General, General Department of Land Transportation, MPWT), Mr. Kim Vothana (Deputy Secretary General, National Committee for Disaster Management), Mr. Koeut Chhe (Vice Governor, Phnom Penh City Hall), Mr. Meas Sam An (Chief, Office of multilateral cooperation, MEF), Mr. Hout Sereymongkol (Chief, Expenditure office 1, National Treasury, MEF), Lorn Chetra (Deputy Office Chief, GDCE), Dr. Sar Chestra (Deputy Secretary General of MAFF), Mr. Ung Chinna (Director, Dep. of Education Quality Assurance, MoEYS), Mr. Pong Pitine (Director of EMIS Department, MoEYS), Mr. Sar Sopheap (Director, Dep. of M&E, MoEYS), Mr. Srey Vuthy (Director, Dep. of Planning and Statistics, MAFF), Mr. Vannarith Chheang (Director, Macroeconomic and Fiscal Policy Department, MEF), Mr. Hav Ratanak (Director of Budget Formulation Department, MEF), Mr. Che Lidin (Director, Ministry of Mines and Energy), Mr. Klok Sam Ang (Director, Dep. of Planning & International Cooperation, MOWRAM), Mr. Muy Monin (Director, Dep. of Finance, MOWRAM), Mr. Mak Mony (Director, Dep. of Planning and Statistics, MAFF), Mr. Kong Phoeun (Director, Dep. of Rural Roads, MRD), Mr. Va Bunnk (Deputy Director, Dep. of Rural Road, MRD), Chan Narith (Deputy Director, Macroeconomic and Fiscal Policy Department, MEF), Mr. Tep Phyorith (Deputy Director, Dep. of Finance, MoEYS), Mr. Hout Sereymongkol (Chief of Expenditure Office 1, MEF), Ly Sophearith (Chief of Audit Program, GDT), Mr. Kim Phally (Chief of Planning and Statistics Office, Dep. of Rural Roads, MRD), Mr. Poliveth Lao (Dep. of Macroeconomics and Fiscal Policy, MEF), Mr. Sopheak Lay (General Department of Budget, MEF), Mr. Ros Borrom, (MEF, General Department of Budget), Mr. Bonnaroth Houl (MEF, General Department of Budget), Mr. Chhuon Samrith (MEF, General Department of Budget), Mr. Natin Patel (MEF, General Department of Budget), Mr. Dy Phanna (Dep. of Finance, MOWRAM), Mr. Sim Piseth (National Economist, Dep. of Macroeconomic and Fiscal Policy, MEF), Mr. Mak Vichetsackada (Dep. of Macroeconomic and Fiscal Policy, MEF), Mr. Pheng Sambo (Dep. of Rural Roads, MRD), Mr. Thou Samnang (Road Data Collection and Management Unit, MPWT), Mr. Gailtan (Road Data Collection and Management Unit, MPWT), Mr. Sreng Sophornreaksmey (General Department of Economic and Public Finance Policy), Mr. Chansophea Chuon (GDCE), Ms. Phanin Hei (Deputy Director General, Phnom Penh port authority), Mr. Hiek Phirun (Deputy Director General, Phnom Penh port authority) and Mr. Hun Sokhalay (Acting head, Dep. of planning/marketing, Phnom Penh port authority). Further contributions were kindly provided by colleagues in development partner organizations: Mr. Jiangfeng Zhang (Director, Environment, Natural Resources and Agriculture Division, South East Asia Department), Ms. Dang Thuy Trang (Environment Specialist, ADB), Mr. Piseth Long (Senior Project Officer, ADB), Mr. Nida Ouk (Senior Project Officer, ADB), Mr. Sideth Muong (Project Officer, Head of Agriculture and Infrastructure sectors, AFD), Mr. Javier del Castillo (Public Financial Management Specialist, EU), Pieter Ypma (Agribusiness manager, CAVAC), Ms. Siv Cheang, (Senior Program officer, Agriculture and Rural Development, JICA), Mr. Koichi Ogawa (Chief Advisor, JICA), Mr. Iwase Hideaki (Project Formulation Advisor, JICA), Mr. Kawano Ryo (Project Formulation Advisor, JICA), Ms. Seng Solady (Program Officer, JICA), Ms. Sang Lee (Agriculture Officer, USAID), Mr. Sophea Ly (Project Management Specialist, Agriculture and Economic Development, USAID), and Mr. Philip Courtnadge (UNDP’s Advisor at the Council of Development of Cambodia). The team is also grateful to Linna Ky, Lyden Kong, Narya Ou, Da Lin, Ravan Chieap, and Chanchamrong Ly for their support during team missions and stakeholder consultations. xii Cambodia: A Public Expenditure Review Executive Summary C ambodia’s and poverty rapid economic reduction have place in an environment characterized by macroeconomic stability and prudent fiscal management. Over the past two decades, Cambodia growth taken likelihood that additional revenue collection gains from the implementation of a very successful Revenue Mobilization Strategy is diminishing; iii) a shift from a heavy reliance on Development Partners (DPs) for the provision of public infrastructure (under a declining grant has been one of the fastest-growing countries in the world component in concessional financing) to self-financing; and has attained most of its Millennium Development and (iv) challenges to public service delivery, including Goals, including that on poverty reduction (World Bank, the provision of quality education for all, the provision of 2017). This has been facilitated by preferential trade public goods in the agriculture sector, and the building treatment, openness to trade and investment flows, and maintenance of a modern transportation network, and macroeconomic stability, against a background all necessary for a successful economic transformation of a pegged exchange rate and a dollarized economy. of Cambodia. Cambodia emerged from the 1998 Asian financial crisis and the 2009 global financial crisis relatively unscathed, The Cambodia Public Expenditure Review (PER) and inflation has been kept at single digits. Following discusses the numerous achievements in fiscal an expansion in the public-sector deficit in 2009 and policy to date and proposes policy options going 2010, authorities have pursued fiscal consolidation and forward, based on both analysis and consultations prudent management of public finances, contributing to with stakeholders. The PER aims to help the Royal overall macroeconomic stability. Government of Cambodia in the effort to overcome the abovementioned challenges and move toward Having recently graduated to become a lower more effective public spending and service delivery, middle-income country, Cambodia is beginning while maintaining macroeconomic stability. Structured to face complex fiscal and public service delivery around the key challenges, the six chapters of the PER challenges. Thanks to sustained growth, GNI per are based on background papers that incorporate capita more than tripled in two decades, from USD 300 feedback from two rounds of consultations with the in 1994 to an estimated USD 1,070 in 2015, the year PER Committee chaired by the Ministry of Economy in which Cambodia became a lower middle-income and Finance (and established by official Prakas on economy. In this setting, Cambodia is facing a series March 18, 2016). In addition, results from sectoral of emerging challenges: i) a decline in DP-related grant discussion papers (Part 2 of this report) have been revenue and rapidly rising public sector wages; ii) the presented publicly at the Budget Discussion workshops Cambodia: A Public Expenditure Review xiii of July 2017. This Executive Summary provides a brief remains at low risk of debt distress. As of 2018, the overview of the analysis and recommendations included Financial Management Information System has been in the PER report. rolled out to all line ministries, which have also adopted program-based budgeting. These efforts are expected to result in increased transparency and a more efficient allocation and execution of public resources. Public finances are expected to remain sustainable, while rising However, quickly rising public sector wages have wage pressures call for improving started putting pressure on the fiscal accounts, and the quality of public spending may eventually result in a suboptimal allocation of public resources. The public sector minimum wage Over the past five years, Cambodia has successfully has been raised from around USD 50 a month in 2012 curbed public sector deficits and made progress in to USD 250 in 2018. As a result, the public wage its Public Financial Management Reform Program bill is expected to amount more than to 8 percent of (PFMRP). Following a peak of 4.4 percent of GDP in GDP in 2018. Although public payroll expenditure as a 2011, public sector deficit (after grants) was reduced in percentage of GDP remains similar to other economies, subsequent years to an estimated 0.2 percent of GDP it represents a high share of government revenues in 2016 (see figure i). This is the result of a significant (more than one-third). Cambodia does not have boost to revenue collection following passage of the too many civil servants; in fact, the number of public Revenue Mobilization Strategy 2014-18, as well as healthcare staff and teachers per population is relatively overall prudent public expenditure. This allowed the low, whereas the ratio of wages to GDP per capita is government to raise deposits (savings) equivalent to 10 significantly higher than in other lower middle-income percent of GDP as of 2016. According to the latest IMF- economies. The surge in public payroll has been driving World Bank joint debt sustainability analysis, Cambodia public expenditure increases across sectors in recent Figure i. While Cambodia has experienced fiscal consolidation in recent years, public payroll pressures are mounting Evolution of public expenditure as % of GDP 25 20 15 10 4.6 5.0 5.7 6.2 6.7 5 3.0 3.4 4.8 4.4 4.3 0 -5 -10 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Wages Provincial expenditure Externally financed capital Other current Locally financed capital Overall balance (excluding grant) Government Revenue Source: World Bank staff. xiv Cambodia: A Public Expenditure Review years, while other recurrent spending has remained alignment with national priorities, while systematically flat in relative terms. In some ministries (Education, monitoring performance data; ii) conducting a functional Agriculture), wages represent the bulk of government and operational review in all ministries, and optimize the spending, while funding to purchases and services allocation of staff, in line with strategic objectives; and needed for the daily operating is limited. iii) strengthening public debt management (including preparing for domestic bond issuances) as financing Going forward, fiscal space is likely to slightly conditions change. It is worth mentioning that the decline, while public finances would remain government has started to move ahead with the Budget sustainable overall. Fiscal space can be defined System Reform Strategy since January 2018. At the as the percentage of revenue available to the central same time, program budgeting can only be effective in government once the most rigid components of improving the allocation of public expenditure towards expenditure have been allocated (payroll, social security the attainment of targets if accompanied with adequate contributions, interest payments, and fixed transfers, and well-motivated human resources. Thus, Public including subnational). Under the baseline scenario, Financial Management Reform and Public Administration fiscal space would decline from 55 percent in 2016 to 47 Reform would need to progress hand in hand. percent by 2023. In the pessimistic scenario, assuming the public sector minimum wage keeps increasing over the next five years to reach USD 400 by 2023, the payroll would be estimated to climb to 10.9 percent Tax incentives could be streamlined of GDP, and fiscal space would drop to 28 percent by to reduce the fiscal cost while 2023. This would significantly constrain the possibility of remaining supportive of efforts to expanding other recurrent spending and government- attract foreign direct investment financed infrastructure. Under the reform scenario, which incorporates some of the recommendations Thanks to a series of tax administration reforms, discussed in this report (refer to section 1.4 for more Cambodia has been able to boost its revenue detail), tax reform (including limitation of tax holidays collection significantly in recent years. Among to six years) would help increase revenue collection, other reforms, since 2008, the General Department of making it affordable to hire teachers to compensate Customs and Excises has successfully rolled out the for existing shortages, while implementation of the Automated System for Custom Data (ASYCUDA) as well PIM framework would allow for an effective increase in as enhanced anti-smuggling measures. More recently, infrastructure spending. While fiscal space under the in the context of the Revenue Mobilization Strategy for reform scenario would remain similar to the baseline, it 2014-18, the General Department of Taxation (GDT) is a more desirable scenario given that it would allow abolished the simplified and estimated tax regimes for filling teacher gaps and increasing public investment and included all small, medium, and large enterprises without incurring a higher deficit. under a self-assessed regime. It has also improved its capacity in tax auditing and taxpayer services, leading To make the most of fiscal policy and available fiscal to better compliance. Thanks to such improvements space, Cambodia needs to continue improving the in tax administration and a broadening of the tax base, allocation and execution of public resources, for Cambodia’s tax revenue collection has increased from increased value for money. Among other measures, 12.7 percent of GDP in 2013 to an estimated 16.1 in this would require: i) making program budgeting an 2016, already above the average for lower middle- effective tool by adopting a Medium-term Budget income economies.1 Framework, consolidating sub-programs, and improving 1  According to revised government figures as of March 2019. They may not exactly correspond to those in the report. Cambodia: A Public Expenditure Review xv However, tax exemptions limit revenue mobilization limiting them to six years; ii) introducing investment gains. In Cambodia, the productivity of the Corporate tax credits as an effective way of lowering firm cost Income Tax (CIT) and Value Added Tax (VAT), defined of acquiring machinery and equipment; iii) replacing as collections to GDP divided by the nominal tax rate, VAT exemptions at customs with import tax credits to has improved in recent years but remains below cross- plug loopholes; iv) streamlining the approval process country averages. This is partly explained by existing for incentives, implementing clearly pre-defined and tax exemptions. Tax expenditure derived from existing objective eligibility criteria and a standard duration; v) custom duty exemptions amounted to around 3.9 reflecting the new incentives scheme in the Tax Code percent of GDP in 2015, with the bulk of exemptions and producing yearly tax expenditure estimates as benefiting Qualified Investment Projects (mostly foreign- an annex to the National Budget for information and owned garment factories). This is larger than the entire transparency purposes. budget to the education sector. Although data quality issues prevented a full assessment of CIT-related tax expenditure, previous analysis by the World Bank (2015) estimated CIT tax expenditure by QIP projects in the Stronger public investment garment and footwear sector to be around 1.8 percent management is needed to step of GDP in 2014. up the provisioning of public infrastructure, traditionally built by Tax incentives in Cambodia could be better linked Development Partners to national objectives and would benefit from streamlining. Currently, incentives in Cambodia are Cambodia has historically met the capital not being used to achieve national objectives such as expenditure goals set in the National Strategic supporting innovation, linking the domestic private sector Development Plans (NSDP) through DP funding, to global value chains, training workers, or focusing which has been on the decline in recent years. on higher value-added sectors. As an overall guiding Public capital expenditure increased from 6.3 percent principle, incentives should be used to encourage the of GDP in 2008 to 10.1 percent of GDP in 2010, then private sector to fund public goods (such as capital progressively decreased to 7 percent of GDP in 2016. formation or infrastructure) and to avoid activities that These trends have been driven mainly by externally are unlikely to generate wider economic and social financed capital, which represents around 75 percent benefits (James, 2014). As Cambodia tries to move up of the total. Domestically financed capital expenditure, the value chain and diversify its exports base, improving historically allocated mainly to smaller-scale road and the investment climate and minimizing the cost of doing hydraulic projects, hovered around 2 percent of GDP business will be critical. This is explicitly acknowledged in 2012-2016. in the Industrial Development Policy, which prioritizes the need to “further strengthen favorable environment Limited progress in the provision of public for investment and doing business by improving the infrastructure points to fragmentation in the capital regulatory framework, rationalizing the provision of project planning, execution, and maintenance. incentives for investment projects” (RGC, 2015b). Although total public capital formation has been substantially higher (an average of 7.5 percent of GDP in In this context, Cambodia could rethink its 2000-15) than in most ASEAN countries, Cambodia still investment incentives mix, limiting tax holidays lags behind other structural peers in terms of extension while moving toward more cost-effective incentives. of the paved road network and quality of transportation A tax incentives reform package could comprise i) infrastructure. Fragmentation of public resources progressively phasing out CIT tax holidays or at least and parallel budget formulation, budget execution, xvi Cambodia: A Public Expenditure Review monitoring and evaluation, as well as financial reporting to provide uniform procedures for the management of undertaken separately for DP-funded and government- project execution at all line ministries; iii) upgrade the funded spending have contributed to this problem. In capacity of MEF, Ministry of Planning, and line ministries addition, in nominal terms, funding for maintenance has to prepare and assess more effective and efficient not kept pace with increases in government-financed investment projects, in connection with the strategic capital. priorities defined in the Budget Strategic Plan (BSP) and Medium-Term Budget Framework (MTBF). Finally, for The Public Investment Management (PIM) framework the government to be successful in complementing DPs for government-funded projects remains weak. in infrastructure provision, the one-year budget cycle Beyond a mention in the Public Financial Management restriction for government-funded projects would need Law, the PIM function in Cambodia is not well defined. to be lifted. To prevent delays in fund disbursement and An assessment following a recognized PIM diagnostic implementation, projects in the short-list approved by methodology (Rajaram et al., 2014) reveals severe the Council of Ministers (around November) should have shortcomings in the legal and institutional framework. specification books ready before January. First, the attribution of roles and responsibilities among key stakeholders is not defined clearly in the Public Finance Law or the existing Public Investment Program (PIP) Sub-decree. For domestically financed investment The government is doubling efforts projects, the absence of a uniform set of guidelines, to increase its stewardship of manuals, and templates for preparation of project the education sector and to keep proposals, appraisal, and monitoring of execution improving access, efficiency, and results in vastly different practices across line ministries. quality of learning outcomes In addition, while the Ministry of Economy and Finance is de facto serving as gatekeeper in the allocation and Attaining a quality education for all is one of use of capital budget resources, independent review Cambodia’s national priorities, and public funding to of project documents and appraisal functions are not the sector has increased significantly in recent years. formally defined. Downstream, project implementation Largely driven by wage hikes, government spending on and operations have shown some progress, with more education has increased significantly in recent years, rigorous procurement and financial controls introduced from an estimate of around 1.6 percent of GDP in 2012 since 2004 in the context of PFMRP, although there is to 2.8 percent of GDP in 2017. Nonetheless, it remains room for further improvement. Finally, Cambodia lacks below the average for low- and lower middle-income effective ex-post evaluation instruments to assess economies (4.1 percent). The introduction of full pilot the efficiency and effectiveness of public investment program budgeting since 2015 has helped improve financing. public expenditure allocation, but challenges persist in terms of budget execution and flow of funds. To address these challenges, three broad and comprehensive measures aimed at strengthening Going forward, authorities could assume further PIM are proposed, namely: i) enact a PIM Sub- stewardship of the sector, since DPs still account for Decree that identifies the stages in the PIM cycle, two-thirds of non-wage expenditure. Development defines institutional arrangements, and introduces partner (DP) and NGO funding, at around 1 percent basic appraisal guidelines, while at the same time helps of GDP in 2010-15, has played a key role in the strengthen the gatekeeping role of MEF in upstream provision of infrastructure. Including DPs, resources decision making; ii) prepare a PIM Standard Appraisal devoted to education are nearly sufficient to cover Manual as well as a PIM Standard Implementation Manual the needs estimated in the Education Strategic Plan, Cambodia: A Public Expenditure Review xvii and allocations are reasonably aligned. However, the options, or grant options would be most optimal for the allocations fall short in higher education and capital Northeastern provinces. spending; authorities could consider increasing the financial autonomy of higher education institutions, Cambodia has a shortage of 40,000 teachers, one- while this would need to come together with more quarter of which could potentially be addressed accountability, to avoid a proliferation of institutions through reallocation. In 2015, about 10 percent of the that do not meet required standards; improving the payroll of the Ministry of Education, Youth and Sports quality of institutions, including in higher education, went to redundant teachers in schools with excess staff, would thus be a pre-requisite and a priority. Any scope while teacher shortages in other schools are equivalent for future increases in education funding should be to 40 percent of the current payroll. Around 43 percent directed toward priority non-wage spending, such as of primary schools with a shortage of teachers have a improving the physical condition of schools, financing school with a surplus of teachers within 5 kilometers. scholarships, and taking over other significant programs Thus, there is scope for reasonable reassignment of that are currently managed by DP organizations. teachers to nearby schools. Alternatively, per diem and incentives could be provided for teachers in the While Cambodia has succeeded in increasing surrounding areas to travel and do shifts in schools with enrollment across levels, secondary school shortages. An effective and long-term solution to poor completion remains the lowest in East Asia. Net teacher allocation would also need to include teacher enrollment rates in primary education are at levels management reform and enforcement of student- similar to those in developed countries. While enrollment teacher ratio norms, preventing teachers from relocating in secondary education has also improved, lower- to schools that are already above the threshold. secondary completion rates (at 48 percent in 2015) remain the lowest in the region. Income explains part Learning outcomes could be improved with the of the existing access gap, as wealthier households introduction of some quality-enhancing measures. are more likely to enroll across all levels, which results Regression analysis of 3rd and 8th grade test scores in in public spending not being pro-poor. In addition, 96 Cambodia shows that, apart from the socioeconomic and 57 percent of non-enrolled children in pre-primary background of the student, absenteeism negatively and primary, respectively, mention they are "too young" affects scores, while those students doing their or they "do not want to go to school" as the reason. homework obtain better marks. This would call for Especially for upper secondary education, Stung Treng, increased parent and community engagement in Oddar Meanchey, and Ratanak Kiri are provinces in school-related meetings, to enhance their awareness, which lack of schools is a significant determinant of as part of the shift towards School-Based Management. children being out of school. These different challenges In addition, teacher behavior in class (e.g., using the require different solutions. First, expanding the coverage blackboard) and understanding of the curriculum and funding of current student grant programs would strongly affects student results. Authorities could help mitigate the opportunity cost of being in school, consider adding a requirement to the current Teacher especially for secondary education. Second, for pre- Policy Action Plan whereby teachers are required to primary, awareness and communication campaigns earn quality-enhancing teaching certificates based on targeting both parents and children would help knowledge of the curriculum and classroom behavior increase enrollment, although this would also need to “best practices”. Technologies to monitor absenteeism be accompanied by building new preschools. Third, could also be introduced. local cost-benefit analysis should determine whether expanding access through construction of more schools The increase in direct funding to schools is or reducing access costs through transport, housing expected to help increase enrollment and improve xviii Cambodia: A Public Expenditure Review learning outcomes, if accompanied by improved transformation, the primary sector remains an essential disbursements and recordkeeping. Small schools in pillar of the Cambodian economy, at about one-quarter less populated and more remote provinces face higher of GDP and nearly half of total employment. per student costs and are found to be in worse condition. Students in small schools and schools with double shifts While public spending in agriculture has increased tend to perform worse. To better respond to different significantly and is well-balanced overall, program school types and needs, Cambodian authorities have fragmentation may be affecting outcomes. Between recently agreed to combine existing sources of funding 2007-2016, government funding to agriculture doubled to schools into a single School Improvement Fund while DP funding tripled, totaling nearly 0.6 percent of (SIF). This would grant increased flexibility in the use of GDP in 2016 (1.8 percent of GDP if irrigation spending is funding at the school level, which is expected to result in included). Overall, the agriculture budget (and its major improved enrollment and learning outcomes, according programs) seems to be well-balanced between wage to international experience. Notably, the shift toward and non-wage recurrent expenditures and is mainly School-Based Management (and increased funding) directed to the provision of public goods. However, would also require a well-functioning authorization and there seems to be underinvestment in a series of core disbursement system, as well as improved financial technical and regulatory functions, including those recordkeeping. Collection of Education Management related to inspection services, SPS compliance, plant Information System (EMIS) indicators will also be and animal health, and agriculture research. Meanwhile, necessary to hold schools accountable and improve the DP funding, at two-thirds of the total, is concentrated in quality of public spending in education. supporting food safety and nutrition, crop production, institutional building, extension services, and fisheries. Funding for livestock and forestry is more limited, and spending levels in these areas fall short of the Despite funding increases, Agriculture Sector Development Plan goals. While there Cambodia underinvests in core are examples of DP-supported projects that are having technical and regulatory functions a positive impact on farm productivity and incomes, essential for a diversified and these impacts have tended to be localized, with limited competitive agriculture spillovers beyond the areas targeted by the projects and limited impact on broader value chains and overall Cambodia’s impressive agriculture performance has sector performance. stalled in recent years, with the fall in commodity prices. Partly supported by rising commodity prices, Despite the scale-up in public investment in recent agricultural GDP growth grew at an average rate of years, less than one-third of Cambodia’s irrigation 4.7 percent between 2006-2012, and the country schemes are fully functional. Driven by DPs (mostly experienced remarkable poverty reduction in this China and Australia), investment in new irrigation period. Cambodia’s laissez faire approach allowed a schemes increased from USD 23 million in 2010 to quick pass-through of international prices to farmers USD 114 million in 2012. However, most of the existing (and farm laborers)—this was accompanied by land irrigation schemes are estimated to be damaged (e.g., expansion and the use of technology, leading to by flooding) or not fully operational. The recent increase improved productivity. However, with the collapse in in government funding for rehabilitation and for operation international commodity prices coupled with adverse and maintenance (O&M) is a welcome step but needs weather events (El Niño), agriculture growth declined to be accompanied by adequate staffing to ensure to an average of less than 1 percent in 2014-16. While effective implementation. Investments in new schemes this has resulted in migration and accelerated structural may still be inefficient, since neither national irrigation Cambodia: A Public Expenditure Review xix and drainage design and construction standards nor terms of quality of transportation infrastructure, which the associated auditing systems have been established does not seem to have improved significantly over the to ensure minimum quality in construction and design. past six years despite significant investments, due to Moreover, integrated water resource management challenges in road maintenance. With 79 percent of (IWRM) activities are underfunded, and the lack of the population living in rural areas as of 2016, it will be an understanding of water availability and lack of important to address the remaining urban-rural gap in transparent rules for managing water usage could pose access to services and rural road accessibility to help a challenge to the sustainable development of water ensure inclusive economic development. resources in Cambodia. The institutional framework for road construction, Going forward, new drivers of agriculture growth will maintenance, and transportation regulation is need to be nurtured through the provision of core fragmented among multiple agencies, and the sector public goods. To this end, an institutional assessment lacks a strategic plan. Cambodia has a Road Law of current capabilities and future needs (both staff and a Law on Land Traffic, but secondary regulation is and funding) in the provision of core public goods underdeveloped, and the distribution of responsibilities (inspection services, SPS compliance, plant and animal is fragmented. The Road Law designates the Ministry health, research) would be critical in the short term. For of Public Works and Transport (MPWT) as responsible irrigation, the focus continues to be in the rehabilitation for the development and maintenance of expressways of existing schemes, while taking into account both and national and provincial roads, while the Ministry of water availability and farm profitability considerations Rural Development (MRD) is responsible for rural roads, and adopting national construction standards for new and cities and municipalities are in charge of urban investments. Finally, there would be a need to elaborate roads. Maintenance of rural roads is very fragmented, and implement a national agriculture research strategy, split between the MRD (for planning, construction, in collaboration with the private sector, universities, and rehabilitation, and emergency maintenance of all rural other relevant actors. roads) and local governments (entrusted with periodic and routine maintenance). While some degree of decentralization is desirable, the intricate and fragmented division of labor and resource allocation often results in To improve the provision of road funding being spread too thinly. An additional level of infrastructure and maintenance, institutional fragmentation and complexity is added with significant institutional the prominent role of DPs in financing and executing fragmentation will need to be investment. Moreover, the transportation sector in addressed Cambodia does not currently have a specific strategy or master plan. Transportation in Cambodia takes place predominantly by road, while accessibility challenges The significant institutional fragmentation is also persist in some rural and remote areas. Cambodia’s reflected in planning and execution processes. Both road network has expanded significantly in recent years, MWPT and MRD have a long-list of investment projects with total length increasing from 46,245 kilometers in which may have originated from requests by provincial 2013 to 61,379 kilometers in 2017. Nonetheless, about departments, sub-national authorities, and politicians 75 percent of the network comprises mostly unpaved and which have different degrees of formalization and rural roads. Only 17 percent of roads were paved as technical rigor. This list includes infrastructure project of 2017, with Cambodia remaining far behind regional proposals that are below a certain size threshold, can be peers. Cambodia also lags other structural peers in executed within a year, and excludes proposals that could xx Cambodia: A Public Expenditure Review be subject to DP financing. Notably, only between 5 to Improving inter-agency coordination, together with 6 percent of the long-list gets financed each year, which project planning, selection, and execution, will be may be due to lack of rigor in planning at the line ministries crucial for enhancing the value-for-money of road (which might submit unrealistic requests for funding) and spending in Cambodia. A series of policy options can which may also indicate systematic underspending, well help address the abovementioned challenges. A first below needs. On the implementation side, the yearly step would be to develop a transportation masterplan budget execution timeline for road projects results in as a guiding sector framework that can help align the significant cost overruns. The first release of funds often efforts of the different actors and prioritize construction occurs around March, which leads inevitably to overlaps and maintenance funding. Related to this, to minimize of works with the rainy season (May to November). the impact of budget fragmentation, the preparation Maintenance during the rainy season may cost about of budget strategic plans by different public agencies 40-50 percent more. should be coordinated as well as aligned with the masterplan. In terms of budget allocation, the recent Transportation spending has fallen in recent emphasis on O&M seems to have led to a significant years, driven by a decline in DP financing of new improvement in road conditions at the national and construction/rehabilitation projects. After peaking provincial level; a survey on rural road conditions in 2012 at around 4.3 percent of GDP, transportation and usage would need to be implemented to better spending steadily fell to an estimated 2.6 percent of GDP determine the allocation of funding going forward. in 2016. Meanwhile, government funding has remained There is also need for a systematic revision of the long- stable relative to GDP (at around 2.2 percent, if sub- list of capital projects by applying public investment national allocations are included). Driven by DP-funded management filtering criteria, keeping only those that spending, capital expenditure declined from 3.4 percent are realistic and in line with national priorities. of GDP at its 2012 peak to 1.4 percent of GDP in 2016, while rebounded to some extent in 2017. Over the past five years, DP-supported road spending has focused on provinces that connect to neighboring countries. Conclusion A recent shift in funding toward O&M has helped In sum, in a context of successful revenue improve national and provincial road conditions mobilization and rising expenditure, Cambodia over the past three years, although more remains to would need to focus in improving the quality of be done. Maintenance spending increased from around public spending and service delivery. In a context 1 percent of GDP to 1.2 percent of GDP between 2014 of rapid economic growth and abundant DP financing, and 2017. National and provincial road conditions have Cambodia has managed to reach lower-middle income improved, from a share of 15 percent of roads in good status under a relatively prudent fiscal policy. In recent condition in 2011 to 17 percent in 2014 and 36 percent years, successful tax administration reform has allowed in 2017. Nonetheless, although there is no inventory or for a significant increase in public sector salaries. While survey of the status of rural roads, anecdotal information public finances are expected to remain sustainable in suggests that the majority are in poor condition. Notably, the coming years, slightly declining fiscal space in the the yearly sub-national budget allocation to maintain all context of further wage increases, coupled with citizen rural roads in the country (representing about three- expectations of higher quality public services, calls for quarters of the total road network, as mentioned above) focusing on improving the quality of public spending is just slightly higher than the budget that Phnom Penh (rather than the quantity). Education, agriculture and City Hall allocates to maintain road infrastructure in the irrigation, and roads are among the areas in which capital city. improvements in expenditure efficiency and public Cambodia: A Public Expenditure Review xxi service delivery can translate into sustained and inclusive projects remains weak. Third, there are daunting economic growth going forward. challenges to program and reform implementation due to limited coordination and communication among There are a series of cross-cutting constraints sector actors (different government agencies, DPs, to the quality of public spending in Cambodia, private sector, universities, etc.). Addressing these identified across the different chapters in this Public issues would require making simultaneous efforts on a Expenditure Review. First, shortcomings in linking series of fronts: Public Financial Management Reform planning, budget, and outcomes are common to all implementation, including FMIS and Public Investment the sectors reviewed, which hinders value for money Management; Public Administration Reform; De- and constrains public service delivery. Second, public concentration and Decentralization; and revamping investment remains outside the budget strategic plans, of multi-actor coordination mechanisms, such as the and it is largely implemented by DPs, while public Technical Working Groups, to make them more agile investment management for government-executed and effective. Cambodia Public Expenditure Review - summary of 20 key selected policy options Challenge / Chapter Short term policy options (1-2 Medium term policy Leading agency years) options (3+ years) I. Value for money • Conduct a functional and opera- • Link wage increases • Ministry of Civil in public spending tional review in all ministries, and and incentives to per- Service remains limited, optimize the allocation of staff, in formance • Line ministries hindering the line with strategic objectives • Fully implement per- and MEF effectiveness of • Make of Program Budgeting an formance-informed fiscal policy effective tool by adopting an MTBF, budgeting consolidating sub-programs and improving alignment with national priorities, while systematically mon- itoring performance data II. Further gains in • Limit CIT tax holidays to 6 years • Replace VAT exemp- • CDC and MEF revenue collection and introduce investment tax cred- tion at customs with are limited by its for capital acquisition import tax credit widespread tax incentives III. Weak legal • Enact a PIM Sub-Decree with • Build capacity for PIM • MEF, in consul- framework and an overall guidelines and prepare at all relevant ministries tation with MP overreliance on PIM appraisal and implementation and agencies • MEF and line DPs persist in the manuals • Create a unique list ministries provision of public • Lift the one-year budget restriction of public investment infrastructure for capital investment, and require projects (including both specification books before January DP and government financed projects) continue in the next page xxii Cambodia: A Public Expenditure Review Challenge / Chapter Short term policy options (1-2 Medium term policy Leading agency years) options (3+ years) IV. Despite increased • Reallocate teachers to nearby • Enforce student-teach- • MoEYS spending in ed- schools, especially for primary er ratios and implement ucation, learning education (43 percent of schools norms that prevent outcomes remain with shortage are within 5 km. of relocations to schools poor schools with excess). When not with excess teachers possible, provide monetary incen- • Introduce the re- tives to travel / do additional shifts quirement of passing • Increase community participation teaching certificates in school-related meetings and en- based on knowledge hance parents’ awareness on the of the curriculum and importance of school attendance classroom behavior and homework “best practices”. Use mobile technologies to better control for teacher absenteeism V. Cambodia • Undertake an institutional assess- • Elaborate a national • MAFF underinvests in ment of current capabilities and agricultural research • MOWRAM core technical future resource needs in inspection strategy addressing and regulatory services, plant and animal health, research priorities as functions essential and water usage well as institutional and for a diversified • Focus on rehabilitating existing financing matters and competitive irrigation schemes that have a • Consider both water agriculture secure water supply, and providing availability and farm capacity building and funding for profitability consider- O&M ations for new invest- ments, and adopt national construction standards VI. There is significant • Conduct a survey of the quality • Build monitoring and • MRD institutional frag- and usage of rural roads evaluation (M&E) mech- • MPWT, in con- mentation in the • Develop a transportation sector anisms to ensure the sultation provision of road strategy or masterplan to align the effectiveness of road infrastructure and efforts of different institutions works and mainte- O&M nance efforts Source: World Bank staff. Cambodia: A Public Expenditure Review xxiii xxiv Cambodia: A Public Expenditure Review PART 1 Government Fiscal Management Cambodia: A Public Expenditure Review 1 1 MAKING THE MOST OF FISCAL POLICY IN CAMBODIA Introduction and motivation: States in the late 1990s contributed to the entry of the macro-fiscal link garment manufacturers that were operating in the O region. Cambodia also became a member of the ASEAN community in 1999 and a member of the World ver the past two decades, Cambodia has Trade Organization (WTO) in 2004, and it has continued sustained an average growth rate of 7.6 to benefit from the “Everything but Arms” preferential percent, resulting in dramatic poverty trade access to the European Union. The adoption of reduction. Cambodia ranks among the most rapidly reforms on companies, investment, land, banking, and growing economies in the world and is expected to join bankruptcy helped foster private sector development the “Olympians of growth,” a group of 13 economies (World Bank, 2017). In 2016, Cambodia ranked 15th in that have grown at an average rate of above 7 percent the world in terms of merchandise trade, at 113 percent a year for 25 years or longer (Ly & Aldaz-Carroll, 2014). of GDP. Growth has been driven by garment manufacturing, agriculture, tourism, and more recently construction Cambodia has been running balance of payments and real estate. Thanks to sustained growth,2 GNI per and fiscal deficits consistently, financed almost capita more than tripled in two decades, from USD 300 entirely by large foreign direct investment (FDI) in 1994 to an estimated USD 1,070 in 2015, the year and foreign aid flows. As a small open economy, in which Cambodia became a lower middle-income Cambodia has seen large FDI inflows into the garment, economy. Poverty incidence has fallen dramatically, construction, and tourism sectors, totaling an average of from 47.8 percent in 2007 to 13.5 percent in 2014, more than ten percent of GDP in recent years. This has according to official estimates. Cambodia has also been helped compensate for wide current account deficits one of the leading countries in reduction of inequality stemming from large imports of both intermediate and in income growth for the bottom 40 percent of the and finished goods (Figure 1, left panel). On the fiscal population. side, authorities have made a substantial effort to curb persistent fiscal deficits, reducing the deficit from 8.8 Cambodia’s outstanding economic performance percent of GDP in 2010 to 2.6 percent of GDP in 2016. has been facilitated by openness to trade and The deficit has been financed mostly by foreign sources improvements in the business environment. The of funding—aid grants as well as loans associated granting of preferential trade treatment by the United mostly with infrastructure projects (Figure 1, right panel). 2  Economic growth has been resilient, with the economy expanding at 5 percent in 1998, the year of the East Asian financial crisis. Economic growth decelerated to 0.1 percent in 2009 but rapidly rebounded and has remained strong, with an average rate of 7.1 percent in 2010-2016. 2 Cambodia: A Public Expenditure Review Cambodia has so far succeeded in maintaining However, despite significant improvements, several macroeconomic stability in the context of a high challenges in human capital formation remain, which degree of dollarization and fast credit growth. could jeopardize growth going forward. Over the past In the late 1980s and early 1990s, Cambodia relied decade, Cambodia has succeeded in raising primary on domestic banks to finance its deficits, which education enrollment to levels comparable to those in resulted in high inflation and encouraged the use high-income economies (Figure 2, left panel). However, of other substitution currencies. The opening up to gross enrollment levels in secondary education, at foreign aid and investors in the 1990s resulted in 45 percent in 2015, remain well below the ASEAN large inflows of U.S. dollars, displacing the Khmer average of 68 percent, and learning outcomes are poor. riel (KHR) as the primary currency in just a few years Public expenditure in education is analyzed in Chapter and bringing extremely fast foreign currency deposit 4 of this PER. In terms of public health care, notable and credit growth.3 These days, the share of foreign improvements have brought very significant declines in currencies in broad money hovers around 80 percent, maternal and infant mortality over the past two decades. and the share of U.S. dollar deposits to total deposits Although out-of-pocket expenditure has also been has remained above 90 percent for the past two reduced significantly in recent years (from 74.2 percent of decades. The Khmer riel is de facto softly pegged total health expenditure in 2014 to 59.4 percent in 2015), to the U.S. dollar, which provides a nominal anchor 4 it remains among the highest in the world, challenging for economic agents and underpins stable prices. equality in access to services (Figure 2, right panel). One It also discourages the public sector from resorting out of three children under the age of five is stunted, to domestic financing and money printing in excess which affects physical and cognitive development. In (World Bank, 2015). Except for the external crises addition, Cambodia does not have a well-developed of 1998 and 2009, inflation has been kept at single social security system and social safety nets, and social digits. assistance spending is just 0.1 percent of GDP, compared to a world average of 1.6 percent of GDP.5 Figure 1. Cambodia’s twin deficits have been financed largely by foreign inflows Current account deficit and FDI, % of GDP Fiscal deficit and foreign aid, % of GDP 12.80 10 11.99 12 10.68 8 10.04 9.14 6 2.5 8 2.6 3.7 4.5 3.5 4 2.7 3.9 2.8 3.4 2.7 4 2 4.8 3.2 4.2 3.2 3.9 2.0 2.8 2.4 1.9 2.3 0 0 -2.4 -2.9 -2.6 -2 -3.8 -2.6 -4 -4 -6.2 -7.6 -6.8 -7.9 -8.8 -6 -8 -6.08 -7.10 -8.22 -7.68 -7.93 -8 -9.36 -8.75 -9.75 -9.85 -12 -10 2012 2013 2014 2015 2016 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Trade Balance Net FDI Current account balance Overall balance Grants Concessional loans (excluding grant) Source: National Bank of Cambodia, Ministry of Economy and Finance (TOFE). 3  Cambodia’s credit to private sector as a percentage of GDP jumped from 2 percent in 1993 to 63 percent in 2015, already above the average for lower middle-income economies. Only Vietnam sustained a faster rate of credit growth in the same period (World Bank, 2017). 4  The currency has been fluctuating between KHR 4,000-4,100 per U.S. dollar since 2011. 5  ASPIRE database. Cambodia: A Public Expenditure Review 3 Figure 2. While primary education is now universal, challenges in pre-primary and secondary education enrollment and overall quality persist Enrolment rates by level Out-of-pocket health expenditure, 2015 120 Yemen Azerbaijan Afghanistan 100 Nigeria Bangladesh Cameroon 80 Pakistan India Sudan 60 Tajikistan Egypt South Sudan 40 Cambodia 59.4 Georgia Morocco 20 Eritrea Venezuela 0 Lower middle-income Cambodia Southeast Asia Lower middle High-income Sierra Leone income Singapore 0 20 40 60 80 100 Pre-primary, % gross Primary, % net Lower secondary, % gross Upper secondary, % gross Out of pocket expenditure (% of total expenditure on helth) Source: Left side: Cambodia information from MoEYS Education Statistics 2015-16, comparator country information and Cambodia pre-primary data from UNESCO Institute of Statistics, 2015. Right side: WHO Global Health Expenditure Database. Shortcomings in the provision of quality state would need to assume increasing responsibility infrastructure also remain a drag on competitiveness for the provision of infrastructure and public services and productivity improvements. While the reliability of to preserve competitiveness and guarantee inclusive electricity supply has improved considerably, only 56 growth. Furthermore, while the pegged exchange percent of the population had access to electricity as of rate under dollarization has enabled macroeconomic 2014 (Figure 3, left panel). In addition, electricity supply stability, it has come at the cost of authorities not being remains costlier than in neighboring countries, hindering able to implement an independent monetary policy. In Cambodia´s competitiveness as a destination for FDI this context, fiscal policy becomes more prominent as inflows. Furthermore, according to the World Economic an instrument (Frankel & Razin, 1987). For fiscal policy Forum, Cambodia seems to have lost its advantage to be used effectively, the allocation and execution of in terms of quality of the road network during the past public spending would need to be improved. decade, which may suggest insufficient maintenance and rehabilitation (Figure 3, right panel). Public The Cambodia Public Expenditure Review (PER) expenditure in agriculture, water, and transportation describes the numerous achievements in fiscal policy infrastructure is analyzed in detail in separate chapters to date and the rising challenges. Going forward, the of this report. Royal Government of Cambodia would need to assume a greater role in closing social gaps, facilitating human For Cambodia as a lower middle-income economy, capital formation, and providing quality infrastructure fiscal policy is expected to become increasingly to maintain economic competitiveness, including the relevant. Despite significant increases in recent years, provision of public goods in agriculture. This represents public expenditure—including a large DP-financed a major challenge and requires efforts to optimize the component—remains low by international standards, allocation of public resources, increase value for money, given Cambodia’s level of GDP per capita (Figure 4). and ensure transparency and accountability in the use As grant and concessional loan financing dwindles, the of public funds, topics that are discussed in detail in this 4 Cambodia: A Public Expenditure Review Figure 3. Access to electricity remains uneven, and Cambodia has lost its edge in road quality Access to electricity (% of population) Quality of roads index (1-7) 120 7 100 6 5 80 4 60 3 40 2 20! 1 0 0 Lower middle- Low-income High-income Cambodia Cambodia Lower middle- ASEAN High-income income income countries OECD 2000 2010 2014 2005-06 2017-18 Source: World Development Indicators and World Economic Forum, Global Competitiveness Index. Figure 4. Public spending remains below international averages Government Expenditure, % of GDP, avg. 2014-16 70 60 50 40 Japan 30 Lao P.D.R. China Myanmar Vietnam +. Thailand 20 Cambodia 2014-16 Cambodia 10 2005-07 0 100 1000 10000 100000 GDP per capita Source: World Economic Outlook, IMF. PER. Part I of this PER discusses overall government fiscal management. Within Part I, the remainder of Recent fiscal policy developments chapter 1 presents recent fiscal developments and After a significant boost in public expenditure discusses future fiscal policy scenarios in Cambodia in 2009, Cambodia has been pursuing fiscal from a sustainability perspective. Chapter 2 focuses consolidation in recent years. Led by an increase in on revenue collection and tax incentives, while Chapter defense and security spending and higher investment, 3 discusses ways to enhance public investment public expenditure rose from 16.1 percent of GDP in management (PIM). Part II of this PER presents the 2008 to 20.5 percent of GDP in 2009 and has remained sectoral expenditure reviews on Education, Agriculture around 22 percent of GDP since then (Table 1). Due to and Irrigation, and Roads. Cambodia: A Public Expenditure Review 5 the increase in expenditure, the general government Externally financed capital expenditure scaled up in the overall fiscal deficit (excluding grants) increased from aftermath of the global financial crisis, from 4.6 percent 2.9 percent of GDP in 2008 to 8.8 percent of GDP in of GDP in 2008 to 7.1 percent of GDP in 2012, then 2010. Authorities responded to the significant increase progressively declined to 4.7 percent of GDP in 2016. in deficit by introducing the ASYCUDA customs system Government-financed capital expenditure has hovered (from 2008) and implementing a Revenue Mobilization around 2 percent of GDP. Strategy (RMS) (2014-18). Combined with an effort to contain expenditure, this has resulted in a progressive Public debt remains low by regional standards, with decline in the general government overall fiscal deficit low risk of distress. Authorities have been avoiding non- from 8.8 percent of GDP in 2010 to an estimated 2.6 concessional external borrowings and domestic bank percent of GDP in 2016. The deficit after grants fell from financing, and most debts are used for the productive 4 percent of GDP in 2010 to 0.2 in 2016. As a result, sectors—particularly for physical infrastructure—while public sector savings surged in recent years, from an the rest are for other priority sectors such as health estimated 4.9 in 2012 to an estimated 12.7 percent of and education. Expenditure on interest payments, GDP in 2017. mainly on foreign concessional loans, thus remains low by international standards, at around 0.5 percent On the revenue side, both direct and indirect tax of GDP in 2012-2016. The 2017 World Bank/IMF Debt revenue have been boosted significantly since 2011. Sustainability Analysis shows that Cambodia’s debt Direct tax collections accelerated from 1.7 percent of distress rating remains low under the baseline scenario. GDP in 2010 to 3.6 percent of GDP in 2016 and have At end-2016, the stock of Cambodia’s external public become more prominent in terms of share of total debt, including arrears, stood at around USD 6.4 billion revenue. Over the same period, indirect tax collections or 31.9 percent of GDP. jumped from 6 to 8.8 percent of GDP, driven by rising excise and VAT collections. The expansion in trade taxes has been slow, hovering around 2.5 percent of GDP despite bustling economic activity, as the country lowers tariffs in compliance with agreements signed with the WTO and ASEAN. Meanwhile, at 2.2 percent of GDP in 2011-2016, non-tax revenue has remained a significant and relatively stable source of collections. A more detailed assessment of revenue is presented in the next chapter. On the expenditure side, the outstanding revenue collection performance has helped finance a significant increase in public sector salaries. The public payroll increased from 4.4 percent of GDP in 2010 to 6.7 percent in 2016 and is expected to keep rising, as the authorities have set a goal for the minimum salary for the public sector to reach one million riel (around USD 250) a month by 2018. Non- wage expenditure increased to over 7 percent of GDP in 2011-14 then fell back to 6.5 percent of GDP in 2016. 6 Cambodia: A Public Expenditure Review Table 1. Cambodia: General Government Operations 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Domestic revenue 12.1 13.4 12.0 13.6 13.6 15.4 15.4 17.7 19.0 19.4 Of which central 11.5 12.6 11.4 12.6 12.3 14.0 14.0 16.3 16.8 17.5 government Tax revenue 9.7 10.6 9.7 10.1 10.2 11.4 11.9 13.8 14.8 14.8 Direct taxes 1.4 1.6 1.7 1.7 1.8 2.3 2.5 2.9 3.4 3.6 Indirect taxes 5.4 6.2 5.5 6.0 6.0 6.8 6.9 8.1 8.7 8.8 Of which: Excises 2.1 2.4 2.0 2.2 2.2 2.5 2.4 3.0 3.3 3.6 (incl. on imports) Domestic 0.2 0.2 0.3 0.4 0.5 0.5 0.6 0.6 0.7 0.8 Import 1.9 2.1 1.7 1.8 1.8 1.9 1.9 2.4 2.6 2.8 VAT (incl. on imports) 2/ 3.3 3.7 3.4 3.6 3.7 4.2 4.4 5.0 5.0 5.1 Domestic 1.4 1.6 1.4 1.6 1.5 1.8 1.9 2.1 1.8 1.9 Import 2.0 2.2 2.0 2.1 2.2 2.4 2.4 2.9 3.2 3.2 Others 0.0 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.3 0.1 Trade taxes 2.9 2.8 2.5 2.5 2.3 2.4 2.4 2.8 2.7 2.4 Provincial revenue 0.6 0.9 0.6 1.0 1.2 1.3 1.3 1.4 2.1 1.9 Nontax revenue 1.7 1.8 1.6 2.2 2.0 2.2 2.0 2.2 1.9 2.5 Capital revenue 0.0 0.2 0.1 0.3 0.1 0.4 0.1 0.3 0.2 0.1 Total expenditure 2/ 3/ 14.0 16.1 20.5 21.9 21.4 22.3 22.1 22.3 21.4 21.9 Current expenditure 8.1 9.0 11.0 11.0 11.4 12.1 12.2 12.8 12.3 13.2 Wages 3.0 3.4 4.8 4.4 4.3 4.6 5.0 5.7 6.2 6.7 Civil administration 2.4 2.3 2.6 2.6 2.5 2.7 3.0 3.6 3.9 4.2 Defense and security 0.6 1.1 2.1 1.8 1.7 1.9 2.0 2.1 2.2 2.5 Nonwage 5.1 5.5 6.3 6.5 7.1 7.5 7.2 7.1 6.2 6.5 of which interest 0.2 0.2 0.2 0.3 0.3 0.5 0.7 0.7 0.3 0.4 payments Provincial expenditure 0.4 0.8 0.7 0.9 0.9 1.0 0.8 1.3 1.9 1.8 Capital expenditure 5.6 6.3 8.8 10.1 9.1 9.2 9.2 8.2 7.1 6.8 Locally financed 1.3 1.7 2.4 2.8 2.6 2.1 2.1 1.9 2.1 2.1 Externally financed 4.3 4.6 6.4 7.3 6.5 7.1 7.1 6.3 5.0 4.7 Continue in the next page Cambodia: A Public Expenditure Review 7 Table 1. Continues 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Expenditure -0.2 -0.2 2.2 -0.6 0.2 -0.2 -0.2 -0.3 0.1 -0.1 Adjustments Current balance (cash) 3.1 3.2 2.5 1.0 1.3 2.0 2.1 4.2 4.4 4.1 Overall balance -2.4 -2.9 -6.2 -8.8 -7.6 -6.8 -7.0 -3.8 -2.6 -2.6 (excluding grant) Overall balance 0.0 0.3 -2.0 -4.0 -4.4 -4.0 -3.1 -1.4 -0.7 -0.2 (including grant) Financing 2.4 2.9 6.2 8.8 7.6 6.8 7.0 3.8 2.6 2.6 Foreign financing (net) 4.7 5.7 6.6 7.1 6.6 6.9 7.1 5.8 4.8 4.4 Of which: Project aid 4.3 5.2 6.1 7.0 6.5 7.0 7.0 6.2 5.1 4.7 o/w grants 1.7 2.7 4.0 4.6 3.0 2.7 3.9 2.3 1.9 2.2 Budget support 0.5 0.7 0.7 0.4 0.4 0.2 0.5 0.0 0.2 0.3 o/w grants 0.3 0.5 0.3 0.3 0.2 0.0 0.0 0.0 0.0 0.1 Debt amortization -0.2 -0.2 -0.3 -0.3 -0.3 -0.3 -0.4 -0.5 -0.4 -0.6 Domestic financing -2.7 -3.6 -0.3 2.4 0.9 -0.8 0.2 -2.2 -2.8 -2.1 Central bank financing -3.6 -2.6 1.6 0.2 0.7 -0.4 -0.1 -2.6 -2.9 -2.4 Nonbank financing 0.9 -1.0 -1.9 2.2 0.2 -0.4 0.3 0.4 0.1 0.2 Outstanding 0.4 0.8 -0.1 -0.7 0.1 0.7 -0.3 0.2 0.5 0.3 operations 4/ Nominal GDP 35,042 41,968 43,057 47,048 52,067 56,681 61,326 67,436 74,423 82,225 (bill. riels) Source: MEF (Budget and TOFE). Notes: 1/ Excludes provincial revenue and expenditure data. 2/ VAT refunds are adjusted. 3/ Adjusting double counting on government tax payments. Total expenditure is based on a mixture of cash and accrual data. 4/ Includes expenditure committed but not yet allocated to the accounts of the government agencies that execute the budget. 8 Cambodia: A Public Expenditure Review Expenditures: prudent spending, infrastructure investment in transport and irrigation, but rising pressures mostly financed by DPs. Expenditure levels by ministries classified as delivering economic services averaged 1.7 percent of GDP in 2004-2008, increasing Public expenditure in social sectors has to an average of 2.6 percent in 2009-2016 (Figure 6, left increased in recent years panel). Most of the variation in economic spending in recent years seems to be linked to large infrastructure Public spending has increased significantly in recent investments financed by DPs, with transport and years, with the increased expenditure levels being irrigation expenditures both receiving significant shares absorbed primarily by the Ministry of Economy and (Figure 6, right panel). The reliance on external funding Finance, defense and security, and social services. is high in Cambodia and is particularly pronounced Average sectoral spending (excluding externally financed in these sectors, where external funding comprises capital) accounted for 8.6 percent of GDP during 2004- around three-quarters of total spending. Meanwhile, 08 but increased to an average of 14 percent of GDP in current expenditure remained stagnant until 2015, when 2009-16. Looking by Ministry, the expenditure average maintenance allocations were increased. for the Ministry of Economy and Finance increased from 0.4 percent of GDP in 2004-08 to 2.4 percent of GDP Budget execution performance is strong at the in 2009-16 (Figure 5). In the aftermath of Cambodia- aggregate level, with small variations between Thailand border instability, expenditure execution by the approved budget allocations and disbursements. The Ministry of Defense and Ministry of Interior - Security most recent PEFA report found small average deviations increased by one percentage point to 2.9 percent of from the budget at the aggregate level, ranging from GDP in 2009-16. Expenditure by social ministries also 2.4 percent to 6.1 percent during 2011-2013 (RGC, expanded, from 3.1 percent of GDP in 2004-08 to 4.2 2015). As shown in Figure 7, this trend has continued. percent of GDP in 2009-16. The large deviation in the economic category in 2014 is related to the punctual implementation of unplanned Increases in spending for ministries associated additional capital spending in energy and roads. with “economic” functions have been driven by Figure 5. Recent expenditure increases have been concentrated in a few key areas Expenditure by Ministry (percent of GDP) 17.5 15 12.5 10 7.5 5 2.5 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016e Defense and Security Public Health Education, Youth and Sport Other social Agriculture, Forestry and Fishery Public Works and Transport Other economic Economy and Finance General administration excl. Economy and Finance Source: World Bank staff calculations based on TOFE (Ministry of Economy and Finance). Cambodia: A Public Expenditure Review 9 Figure 6. Spending in both economic and social services has been on the rise Govt Expenditure, Govt expenditure breakdown, by aggregate functional categories (% of GDP) economic subcategories (% of GDP) 18% 3.5% 16% 3.0% 14% 2.5% 12% 10% 2.0% 8% 1.5% 6% 1.0% 4% 0.5% 2% 0% 0.0% 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 General Administration and Judiciary Social Services Rural Development Water Resources and Meteorology Defence and Security Rural Development Agriculture Other Economic Services Economic Services Environment Transport Source: World Bank staff calculations based on TOFE (Ministry of Economy and Finance). Figure 7. Deviation between budgeted and executed current expenditure, by Ministry sub-categories Average Deviation in absolute value Deviation by year (% deviation from budget) 25% 70% 60% 20% 50% 40% 15% 30% 20% 10% 10% 5% 0% -10% 0% -20% Defense Social Economic General 2010 2011 2012 2013 2014 2015 2016 and Security Administration excl. MEF General Administration excl. MEF Social Avg. Deviation 2010-12 Avg. Deviation 2013-16 Defense and Security Economic Source: World Bank staff calculations based on TOFE (Ministry of Economy and Finance). With the introduction of program budgeting, feed into a top-down national MTBF, which constitutes Cambodia has made significant progress towards the basis for budget formulation. The BSPs attempt strengthening the links between planning and to provide the link between the government priorities budgeting, while further efforts are required to under the National Strategic Development Plan, the pursue a fully functioning Medium-term Budget sector plans, and the annual budget. Efforts also have Framework (MTBF). Since 2015, most line ministries been made to gradually improve costing and set more are required to submit to the MEF 3-year rolling bottom- realistic outcome and output targets. Another positive up strategic budget plans (BSP), which ultimately would feature is the shift towards results-based management, 10 Cambodia: A Public Expenditure Review with program managers becoming responsible for DPs and Civil Society Organizations have been held at the implementation of programs and sub-programs, the parliament.  In 2016, the Ministry of Economy and including in terms of administering funding (previously Finance started to produce a yearly Economic Monitor done centrally). However, progress remains uneven as well as a mid-year budget review. and further improvements are needed before the BSPs can be considered fully a reflection. Namely, programs The lack of comprehensive capital budgeting remain in many cases a reflect of administrative divisions somewhat distorts the picture, preventing a at the ministry, estimates (and fund requests) are often complete measurement of execution performance. unrealistic, and the link between inputs and outcomes Capital budget appropriations by ministry were only remains weak. Information about DP-funded projects is introduced in 2015. Most ministries still execute capital often missing in BSPs, which also results in the MTBF being spending on an ad hoc basis without an initial capital incomplete, as it reflects just part of total public spending. budget, and the major capital expenditure ministries routinely execute far above 100 percent due to ad hoc There has been some improvement in the allocations. This weakens the ability to assess execution publication of planning and budget information, against budget. However, estimates of overall capital while transparency in the decision-making process execution based on Treasury reports indicate that could be further enhanced. Planning and budgeting execution rates have been strong, increasing from an information sharing has improved to some extent through average of 86.7 percent in 2011-2014 to 92.5 percent regular publication of monthly fiscal reports (TOFE) and for 2015-2016. Budget composition and predictability the publication of an annual budget “in brief”. Starting may improve further as the recently introduced program in 2015, yearly budget discussion workshops open to budgeting scheme is implemented fully (see Box 1). Box 1 Budget integration underpins spending efficiency Cambodia has made significant progress in reforming public expenditure policy and public finances after embarking on a Public Financial Management Reform Program (PFMRP) in 2005. Out of the four public financial management areas signaled in the 2011 Integrated Fiduciary Assessment and Public Expenditure Review (IFAPER), progress has been achieved in three of them: use of the government chart of accounts for DP-financed projects, adoption of the Financial Management Information System (FMIS) by the Ministry of Economy and Finance, and alignment of the subnational reform agenda with PFMRP. However, no significant progress has been made in the integration of externally financed (DP-funded) and domestically financed (government-funded) budgets. Budget integration is crucial. While the disbursement profile, especially for capital spending, has been improved significantly thanks to better budget execution processes and sectoral capital budget appropriations, it appears that numerous weaknesses continue to hinder the management of capital investment: (i) parallel budget formulation and implementation, (ii) weak reporting systems for domestically and externally financed budgets, (iii) insufficient policy-budget linkages, and (iv) still incipient Medium-term Budget Framework (MTBF) and unrefined budget strategic plans (BSP). The lack of detailed reporting on implementation of externally funded projects, which constitute three-quarters of public investments, is identified by the 2015 Public Expenditure and Financial Accountability (PEFA) report as a weakness (RGC, 2015). Extent of unreported government operations (PI-7) is scored as “C,” and Financial info provided by DPs for budgeting and reporting on projects/programs (D-2) gets a “D+.” The PEFA report also pointed out the need for strong policy- Cambodia: A Public Expenditure Review 11 Box 1 budget links in resource allocation over the medium term to achieve high-level policy objectives in public service delivery (“C” score for dimensions (iii) and (iv) under PI-12). It also appears that sufficient attention has not been paid to improving the capital expenditure framework and the capacity to manage domestically financed public investment. Going forward, the successful application of the new, seven-segment uniform account code structure (UACS) for both externally and domestically financed spending through the FMIS is expected to facilitate budget integration. As emphasized in the PEFA report, monitoring of budget allocations and their use has been undermined historically by the lack of a unified chart of accounts reflecting classifications based on sector, function, and program ( “C” score under PI-5: Classification of the budget). The seven-segment classification system integrates the budget and accounting functions into a uniform single structure. Each segment provides a different aspect of the expense and revenue and together provides the overall framework reflecting the needs of all the users of financial data provided by the government, ranging from the parliament, the National Audit Authority, ministries, and sub-national governments to citizens and DP agencies. The FMIS will enable controls of aggregate spending and deficits, while improved MTBF allows prioritization of expenditures across policies, programs, and projects for allocative efficiency and equity. The steps to advance budget integration include: (i) application of all segments of UNACS, (ii) rolling out FMIS to line ministries to enable Project Implementing Units to report DP-financed programs, and (iii) ensuring uniform application of the revised Chart of Accounts to improve accuracy in recording and reporting transactions. Recent expenditure consolidation efforts represents a high share of government revenues. At are now facing pressures from rising an estimated 6.7 percent of GDP in 2016, Cambodia’s wages public wage bill was above the average for Asia and the Pacific but similar to the level of low-income economies Although public expenditure has been contained and below that of emerging economies. Nonetheless, in recent years, Cambodia faces increasing wage as a share of government revenue (34.5 percent in pressures. The public sector minimum wage was 2016), it was already significantly higher than in most raised from around USD 50 a month in 2012 to USD countries (Table 2). 145 in 2016, and the government has announced it will reach USD 250 (or more than 8 percent of GDP) Compared to other countries, Cambodia does by 2018. Minimum wage increases have brought an not have a large number of civil servants, but the increase in the overall public payroll, from 4.3 percent of ratio of wages to GDP per capita is significantly GDP in 2011 to 6.7 percent in 2016 (Figure 8, left panel). higher. Cambodia does not seem to have a high ratio Wage spending as a share of government expenditure of civil servants to total population (Figure 9, left panel). increased from around 20 percent in 2010 to more than However, Cambodia’s ratio of civil administration wage 30 percent in 2016 (Figure 8, right panel). to per capita GDP of 3.1 is the highest among surveyed countries (Figure 9, right panel), which points to better Although public payroll expenditure as a percentage remuneration of an average government employee in of GDP remains similar to other economies, it relation to living standards. 12 Cambodia: A Public Expenditure Review Figure 8. The share of public sector wage to total spending has risen significantly in recent years Government expenditure ratios 25 350% 300% % of government expenditure 20 250% Percent of GDP 15 200% 150% 10 100% 5 50% 0 0% 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Other current Provincial expenditure Wages Locally financed capital, % government expenditure Externally financed capital Locally financed capital Wages, % government expenditure Source: World Bank staff calculations using TOFE (Ministry of Economy and Finance). Table 2. The public-sector wage bill represents a higher proportion of government revenues than in other countries Government % of GDP % of government % of government compensation expenditures revenues Cambodia 6.7 30.6 34.5 Asia and Pacific 6.6 27.5 22.9 Low-income 7.0 25.8 26.8 Middle-income 8.7 31.8 28.2 European Union 9.9 25.4 25.3 High-income 10.4 28.0 25.9 Source: Cambodia’s 2016 TOFE and IMF, “Evaluating Government Employment and Compensation.” As examples, the education sector and health sector standards (Figure 10, left panel). Similarly, health sector both have small shares of employment relative to employment as a share of total population is very small the total population. Education sector employment at 0.2 health personnel per 100 population (Figure 10, as a share of total population is relatively small at 0.8 right panel). However, unlike for education, the wage bill education personnel per 100 people—like that of low- of the Ministry of Health as a share of total spending income countries. Interestingly, if development partner is well below average if the DP-funded component is (DP)-funded expenditure is excluded, Cambodia’s excluded, probably due to the very limited number of education sector wage bill as a share of total Ministry staff in the sector. spending (80.5 percent) is high by international Cambodia: A Public Expenditure Review 13 Figure 9. Cambodia has a small proportion of civil servants but a higher ratio of public wages to GDP per capita Civil servants in % of total population Ratio of average public administration wage to GDP per capita 12 3.5 3,1 10 3 2.5 8 2 6 1.5 4 2.9 1 2 0.5 0 0 Cambodia 1/ Africa Low-Income Asia & Pacific Mid. East & C. Asia W. Hemisphere Middle-Income High-Income European Union Europe Mid. East & C. Asia High-income Africa European Union Asia & Pacific Europe W. Hemisphere Mid-income Low-income Cambodia Source: Cambodia's 2016 budget and IMF, “Evaluating Government Employment and Compensation”. Note: 1/ Assuming security and defense staff is the same as that of civilian personnel. Figure 10. Employment as a percentage of total population is relatively small in both education and health, while the payroll share in sectorspending is more divergent between the two sectors Public education sector employment Public health sector employment % of health spending in education budget 70 % of wage spending in education budget 3 90 3 Public education staff, % of population Public health staff, % of population 80 60 2.5 2.5 70 50 2 60 2 50 40 1.5 1.5 40 30 1 30 1 20 20 0.5 0.5 10 10 0 0 0 0 Cambodia Low-Income Africa Asia & Pacific Mid. East & C. Asia W. Hemisphere Middle-Income High-Income Europe Cambodia Low-Income Asia & Pacific W. Hemisphere Mid. East & C. Asia Africa Middle-Income High-Income Europe % of population % wage spending in budget Source: Cambodia's 2016 MEF Treasury report and IMF, “Evaluating Government Employment and Compensation.” 14 Cambodia: A Public Expenditure Review Meanwhile, other recurrent spending and Capital expenditure in Cambodia has increased capital spending have not increased in at a slower pace since 2010, while declining as a real terms percentage of GDP. Total public capital expenditure increased from 6.3 percent of GDP in 2008 to 10.1 Within other current spending, interest payments percent of GDP in 2010 then progressively decreased to account for most of the increase in real terms, 6.8 percent of GDP in 2016 (Figure 11, left panel). These while other categories have remained flat. Interest trends have been driven mainly by externally financed payments account for part of the recent increase in capital, which represents around 75 percent of the total. other current expenditure, a product of the increase in Locally financed capital expenditure, which historically concessional borrowing observed in the aftermath of has been allocated mainly to smaller-scale road and the 2009 crisis. However, at an average of 0.3 percent hydraulic projects, hovered around 2 percent of GDP of GDP in 2009-2012 and 0.5 percent in 2013-2016, in 2012-2016. Budget allocations in 2017 increased interest payments remain modest. During 2010-16, significantly to 2.4 percent of GDP. Meanwhile, in public purchases 6 stayed around 1.6-1.8 percent nominal terms, funding for maintenance has not kept of GDP, external services (mainly maintenance and pace with increases in government-financed capital repairs) at 1 percent of GDP, and other services (mainly (Figure 11, right panel). transportation, advertisement and telecommunications) at 0.6 percent of GDP. Transfers to provincial authorities, Nonetheless, Cambodia has been able to meet administrative institutions, and cultural entities remained the capital expenditure goals set in the National slightly above 1 percent of GDP in the same period. Strategic Development Plans (NSDP). As shown Meanwhile, expenditure classified as social assistance in Table 3, both domestically and externally financed increased to 1.5 percent of GDP in 2011 then fell public investment targets for the NSDP update (2009- back to the 1.1 percent level observed in 2010, a level 2013) were surpassed by a wide margin. The targets set maintained through 2016. for NSDP years 2014 and 2015 were also surpassed, Figure 11. Growth in capital expenditure has slowed since 2010, mainly due to a slowdown in externally financed capital, and funding for maintenance has not kept pace with domestically financed capital spending Capital Expenditure (% of GDP) Domestically financed investment and maintenance (in bn. riel) 12 2,000 1,800 10 1,600 8 1,400 1,200 6 1,000 800 4 600 2 400 200 0 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2007 2008 2009 2010 2011 2012 2013 2014 2015 1016 Locally financed Externally financed Maintenence Capital Source: World Bank staff calculations using TOFE (Ministry of Economy and Finance). 6  This category includes maintenance supplies, office supplies, food and agricultural products, clothing and distinctions, small tools, material, furniture and equipment, water and energy, sanitary and health products, and other supplies. Cambodia: A Public Expenditure Review 15 Table 3. Cambodia has nonetheless been meeting its capital expenditure goals in the NSDP, particularly for domestically financed investment Capital Expenditure NSDP Actual NSDP Actual NSDP Actual + (mill USD) Update projected Year 2009-2013 2009-2013 2014-2015 2014-2015e 2014-2018 2014-2018 Public Capital Invest- 4,339.0 6,098.9 2,611.8 2,684.4 7,671.8 7,524.5 ment o/w Domestically 1,020.3 1,474.9 643.0 686.4 1,888.6 2,229.3 financed o/w Externally financed 3,318.6 4,624.0 1,968.8 1,998.0 5,783.2 5,295.1 Source: National Strategic Development Plans and TOFE (Ministry of Economy and Finance). Figure 12. Cambodia’s relatively large infrastructure investment has been supported by DPs Net investment in nonfinancial assets Net ODA received per capita, (% of GDP), avg. 2010-2015 average 2010-2015 (current US$) 12 80 70 10 8.4 60 51.7 8 50 6 40 30 4 20 2 10 0 0 Upper middle-income Middle-income Indonesia Lower middle-income Philippines Thailand Malaysia Vietnam Low-income Cambodia Lao PDR Thailand Indonesia Philippines Upper middle-income Vietnam Cambodia Low-income Lao PDR Source: WDI data, World Bank staff calculations using TOFE (Ministry of Economy and Finance). but projections for the entire NSDP period indicate that panel). Similar to Lao PDR and Vietnam, high per capita total public capital investment could fall short of the development assistance to Cambodia helps finance total target, despite domestically financed investment greater public capital spending (Figure 12, right panel). exceeding targets. Enhanced DP coordination, avoiding a piecemeal approach to project selection, and having adequate Net yearly infrastructure investment in Cambodia is operations and maintenance budgets are important for high relative to neighboring countries, partly due to ensuring that large public capital expenditure translates the large reliance on DPs. Cambodia ranks second into good quality infrastructure. The next section after Lao PDR among ASEAN countries in terms of net discusses externally financed expenditure allocations, investment of non-financial assets, at an average of while chapter 3 elaborates further on public investment 8.4 percent of GDP during 2010-2015 (Figure 12, left management. 16 Cambodia: A Public Expenditure Review A look at Development Partner financing administration and judiciary (core government) has been declining steadily as a share of GDP. Similarly, following After peaking in 2011-12, external aid and an increase to an average of 2.8 percent of GDP in concessional finance flows as a share of GDP have 2011-12, DP spending in social services declined to 1.9 been on the decline. Taken together, DP and NGO percent of GDP in 2013-15. Funding to health and social disbursements doubled from USD 713 million in 2006 assistance has declined, while funding to education has to USD 1,499 million in 2012 then fell to USD 1,269 been maintained overall. million in 2016. As a percentage of GDP, they declined from 11 percent of GDP in 2011 to 6.4 percent of GDP In terms of types of financing, there has been a shift in 2016 (Figure 13, left panel). Over 2008-2014, NGO toward loan financing in recent years. Aid grants disbursements accounted for a relatively stable share at received by Cambodia declined from around 6 percent around 14 percent of the yearly total. of GDP in 2009 to 3 percent of GDP in 2015, the year in which the country graduated from low-income status The functional composition of DP financing changed (Figure 14, left panel), and some bilateral and multilateral markedly in the aftermath of the 2009 crisis, with a DPs have been gradually reducing their grant funding clear shift toward economic services. Most of the shares. Until 2010, the loan share in total financing additional funding during the peak years of 2011-12 was stable at around one-third of total DP financing. came from an increase in spending in transportation It then increased to close to half of total financing in and agriculture. The share held by economic services just two years, 2011 and 2012. During the period expanded from 41.4 percent in 2008-10 to 52 of rapid increase in DP funding, China emerged as a percent in 2011-12 then to 56.3 percent in 2013-15 major source of financing and now represents around (Figure 13, right panel; Table 4). DP-financed capital 30 percent of total DP funding. Since most of Chinese spending in economic sectors has helped complement aid is in the form of concessional loans (with grants modest government-financed allocations to these representing just 0.5 percent of total Chinese funding in sectors (averaging around 1.6 percent of GDP) (see 2000-15), this has contributed to the observed shift in Figure 6, right panel). Meanwhile, funding to general the profile of aid. Notably, loan funding is prevalent in the Figure 13. Total Development Partner inflows have declined since 2011 Total DP Funding (including NGO) DP funding (excl NGOs) composition by major functional categories (Mill USD) 1,600,000 14% 1,400 1,400,000 12% 1,200 1,200,000 10% 1,000 1,000,000 8% 800 800,000 6% 600 600,000 4% 400 400,000 200,000 2% 200 0 0% 0 2008 2009 2010 2011 2012 2013 2014 2015 2016e 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Core Government Rural Development Other Total Disbursement (mill USD) Total Disbursement (% GDP) Social Services Economic Services Source: Calculations by World Bank staff based on Council for the Development of Cambodia (CDC) database. Cambodia: A Public Expenditure Review 17 Table 4. DP financing has shifted in recent years toward economic services Functional category Avg shares 2008-2010 Avg shares 2011-2012 Avg shares 2013-2015 share of DP % GDP share of DP % GDP share of DP % GDP funding funding funding Core Government 13.5% 1.1% 8.3% 0.8% 7.8% 0.6% General Administration & Judi- 13.5% 1.1% 8.3% 0.8% 7.8% 0.6% ciary Defense 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Security (Interior) 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Economic Services 41.4% 3.4% 52.0% 4.8% 56.3% 4.1% Agriculture 7.9% 0.6% 12.6% 1.2% 15.6% 1.1% Transport 20.0% 1.6% 26.1% 2.4% 27.1% 2.0% Other Economic Services 13.5% 1.1% 13.2% 1.2% 13.6% 1.0% Environment, incl climate 2.2% 0.2% 1.2% 0.1% 2.7% 0.2% change Rural Development 6.8% 0.6% 4.2% 0.4% 5.0% 0.4% Social Services 29.6% 2.4% 29.9% 2.8% 26.1% 1.9% Health 19.2% 1.6% 14.9% 1.4% 13.9% 1.0% Education 7.6% 0.6% 7.5% 0.7% 8.4% 0.6% Other Social Services 2.8% 0.2% 7.5% 0.7% 3.8% 0.3% Other 3.3% 0.3% 1.8% 0.2% 0.5% 0.0% Emergency & Food Aid 1.6% 0.1% 2.0% 0.2% 1.6% 0.1% Budget & BOP support 1.6% 0.1% 0.6% 0.1% 0.0% 0.0% Total Calculated 100.0% 8.2% 100.0% 9.3% 100.0% 7.2% Source: Calculations by World Bank staff based on CDC database. economic service sectors and particularly in transport expansions in economic function projects, external (Figure 14, right panel), while the social sectors have funding in capital expenditure, and loan share in external been relying mainly on grants. financing. While investment project funding makes up between 80-90 percent of total economic services In terms of types of projects, investment project expenditures (and 100 percent in the case of transport funding is clearly dominant in the economic services projects), it only represents between 25-45 percent of categories, while technical assistance dominates in social services sector funding. Technical assistance still the social sectors. Investment Project Assistance (IPA) accounts for nearly one-third of overall DP financing and as a share of total DP funding increased from less than 50 is concentrated in the provision of general administration, percent before 2008 to an average of around 65 percent judiciary, and social services. in 2011-15. This is related to the abovementioned 18 Cambodia: A Public Expenditure Review Figure 14. Cambodia has seen a decline in DP assistance and a shift in DP funding composition from grant to loan financing, associated with an increase in economic sector spending Terms of external financing (% of GDP) Average sector DP funding, loan and grant, average 2010-15 (% of GDP) 14% 2.5% 12% 2% 10% 1.5% 8% 6% 1% 4% 0.5% 2% 0% 0% General Agriculture Transport Health Education 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016e Admin and Judiciary Grant Loan NGO Grant Loan Source: Calculations by World Bank staff based on CDC database. Toward a more effective and in 2012-15. Fiscal expansion is expected to be driven sustainable fiscal policy in largely by the rising wage bill, as discussed earlier. Cambodia The revenue collection forecasts are derived from projections of each of the main revenue components (direct, indirect, and trade taxes as well as non-tax Medium-term fiscal framework revenue), computed using their past trends observed projections: creating fiscal space for during 2007-16. It is assumed that no additional taxes reforms or changes in tax rates are introduced over the coming years. The underlying assumption is that both direct This analysis compares three possible revenue and VAT tax elasticities with respect to GDP remain collection scenarios as Cambodia continues constant at 1.05 over the whole period, while in the case its expansionary fiscal policy. Following fiscal of taxes on specific goods and services, the elasticity consolidation in recent years, the expansionary fiscal is assumed to be equal to 1 (constant proportion of policy started in 2017 is likely to be sustained in GDP). Finally, international trade taxes are assumed the medium term. Against this background, three to grow at 8 percent in nominal terms, thus showing potential scenarios are presented here. In addition to a slight decline in terms of GDP during the period to the baseline scenario, alternative scenarios informed capture tariff decreases under ASEAN Free Trade by conversations with authorities and feasible reforms Area agreements. In this baseline scenario, revenue described in this PER have been analyzed. collection would remain at around 18.5 percent of GDP over the forecast period. The returns to improvements The baseline scenario draws on the analysis of in tax administration would likely decrease, as some of past revenue trends to forecast future collection. the low-hanging fruits would have already been picked. Under this “business as usual” scenario, total public expenditure is projected to average 23.7 percent of The reform scenario assumes that the Government GDP in 2017-23, compared to 22.0 percent of GDP implements a tax reform that affects both tax Cambodia: A Public Expenditure Review 19 elasticity and tax buoyancy. Tax elasticity of direct to cover shortages in primary and secondary education, and indirect taxes is assumed to grow from 1.1 to 1.2 as discussed in Chapter 4 of this PER. This would lead by 2019 and further to 1.25 during 2019-2023, in the to a one-time rise in the public payroll of 0.8 percentage context of implementation of a new Revenue Mobilization points in 2019, reaching 9.2 percent of GDP. In Strategy. The following proposed reforms are expected 7 addition, passing of the PIM reform (see Chapter 3 of to have one-time effects in 2019: limitation on CIT tax this PER) would allow the RGC to raise government- holidays to six years (direct tax collections rise by up financed capital spending to 4 percent of GDP by 2030, to 0.4 percentage points of GDP); introduction of input helping to compensate for the decline in DP-financed tax credits instead of exemptions at the border (VAT investment. As a result, public expenditures reach 25.1 tax collections increase 0.3 percentage points of GDP); percent of GDP in 2023. However, due to a reduction in and full enforcement of the property tax (“other tax” tax incentives, the primary deficit under this scenario is collections rise 0.4 percentage points of GDP). Under still lower than under the baseline scenario. this scenario, revenue collection in terms of GDP rises around 1.3 percentage points on average relative to the The pessimist scenario assumes that wages baseline during 2017-2023, reaching 20.6 percent of continue rising over the next five years. Assuming GDP in 2023 (Table 5). Neither taxes on specific goods the minimum wage of the public sector keeps increasing and services nor international trade taxes are assumed over the next five years, reaching USD 400 by 2023, the to change with respect to the baseline scenario. payroll is estimated to climb to 10.9 percent of GDP. In this scenario, to compensate for wage increases, Under the reform scenario, authorities can afford government-financed capital expenditure is crowded to hire new teachers and increase capital spending out and does not keep increasing. As a result, public while maintaining fiscal sustainability. The reform expenditure rises from 22.6 percent of GDP in 2017 to scenario assumes that authorities recruit new teachers 26.6 percent in 2023. Since the revenues do not rise at Table 5. Revenue collection forecasts under alternative scenarios (% of GDP) 2017p 2018p 2019p 2020p 2021p 2022p 2023p Baseline scenario 18.5 % 18.5 % 18.5 % 18.5 % 18.5 % 18.5 % 18.5 % Pessimist scenario 18.5 % 18.5 % 18.5 % 18.5 % 18.5 % 18.5 % 18.5 % Reform scenario 18.6% 18.7% 19.8% 20.0% 20.2% 20.4% 20.6% Source: World Bank staff calculations based on official figures (TOFE). Note: p = projection. Table 6. Evolution of primary balance under alternative scenarios (% of GDP) 2017p 2018p 2019p 2020p 2021p 2022p 2023p Baseline scenario -3.7 % -5.2 % -5.4 % -5.0 % -4.8 % -4.5 % -4.3 % Pessimist scenario -3.6% -4.6% -5.7% -6.2% -6.7% -6.9% -7.2% Reform scenario -3.6% -4.9% -5.0% -4.7% -4.5% -4.1% -3.9% Source: World Bank staff calculations based on official figures (TOFE). Note: p = projection. 7  Regarding the plausibility of tax elasticity assumptions, it should be noted that for the reform scenario, they would imply a reduction in informality of approximately 7 percentage points by 2023, assuming that the shadow economy equals 48.7 percent in 2017. This figure is based on Buehn and Schneider (2011). 20 Cambodia: A Public Expenditure Review the same rate, the fiscal accounts strongly deteriorate. Under both the baseline and the reform scenarios, The primary deficit is forecast at 7.2 percent of GDP in fiscal space would slightly decrease, while remaining 2023, more than 3 percentage points of GDP higher than adequate, and public-sector debt is expected to under the baseline scenario (Table 6). Projections under be sustainable. Fiscal space can be conceptualized the pessimist scenario do not present any difference in as the percentage of revenue available to the central terms of tax collection with respect to the baseline. government once the most rigid components of expenditure have been allocated: payroll, social security In all three scenarios, the public expenditure outlook contributions, interest payments, and fixed transfers, is driven mainly by the rising revenue bill. Under including subnational (Gonzalez, et al., 2012). According the baseline scenario, it is assumed that the wage to this definition, fiscal space in Cambodia has remained bill increases until 2018 as announced by authorities, ample at an average of 60 percent of revenue in 2009- reaching 8.4 percent of GDP in 2018. No significant 2016, while it will experience some decline in the coming wage increases would be enacted from 2019 onward, years, mostly driven by public wages (Figure 15, left which would result in the public payroll easing in relative panel). Fiscal space is expected to average 46 percent terms until it falls back to 7.8 percent of GDP in 2022 during 2017-2023 under the baseline scenario and (Table 7). The domestically financed component of would be just slightly lower under the reform scenario.8 public capital expenditure is assumed to grow steadily Under the pessimist scenario, fiscal space would from 2.6 percent of GDP in 2017 to 3.5 percent of deteriorate significantly, dropping below 30 percent GDP in 2017, as authorities try to compensate for the by 2023. In terms of sustainability, public sector debt expected fall in externally financed capital. However, (mostly foreign) would increase in all three scenarios, in should there be no progress in the implementation of the context of rising salaries. Under the baseline and a Public Investment Management (PIM) framework, the reform scenarios, the debt stock would peak at around prospects for further effective increases in government- 40.5 and 39 percent of GDP, respectively, in 2022, then financed capital would remain limited. decline mildly. Under the pessimist scenario, it would Table 7. Public expenditure forecasts under alternative scenarios (% of GDP) 2017p 2018p 2019p 2020p 2021p 2022p 2023p Baseline scenario 22.6 % 24.1 % 24.4 % 24.1 % 23.9 % 23.7 % 23.5 % wages 7.6 % 8.4 % 8.2 % 8.0 % 7.8 % 7.8 % 7.8 % domestic cap. exp. 2.6 % 3.0 % 3.1 % 3.3 % 3.4 % 3.5 % 3.5 % Pessimist scenario 22.6% 23.7% 24.8% 25.4% 26.0% 26.3% 26.6% wages 7.6 % 8.4 % 8.9 % 9.4 % 9.9 % 10.4 % 10.9 % domestic cap. exp. 2.6 % 2.6 % 2.6 % 2.6 % 2.6 % 2.6 % 2.6 % Reform scenario 22.6 % 24.1 % 25.3 % 25.3 % 25.3 % 25.2 % 25.1 % wages 7.6 % 8.4 % 9.2 % 9.2 % 9.2 % 9.2 % 9.2 % domestic cap. exp. 2.6 % 3.0 % 3.1 % 3.3 % 3.4 % 3.6 % 3.7 % Source: World Bank staff calculations based on official figures (TOFE). Note: p = projection. 8  Under the reform scenario, the fiscal space shrinks even assuming a significant boost in revenue collection, since the rigid component of public expenditure (mainly teacher wages) rises faster. Cambodia: A Public Expenditure Review 21 Figure 15. In the pessimist scenario, fiscal space and debt sustainability would deteriorate significantly Fiscal Space Public sector debt, % GDP 60% 48% 46% 50% 46.86 44% 40% 42% 39.41 30% 40% 38% 20% 36% 37.62 10% 34% 0% 32% 2017p 2018p 2019p 2020p 2021p 2022p 2023p 2017p 2018p 2019p 2020p 2021p 2022p 2023p Base Scenario Pessimist Scenario Reform Source: World Bank staff calculations based on official figures (TOFE). Note: p = projection. continue rising to more than 46 percent of GDP during economy constraints that prevent them from saving in the forecasted period and may become unsustainable periods of strong economic growth (Alesina, Campante, (Figure 15, right panel). & Tabellini, 2008), whereas imperfections in international capital markets preclude them from accessing financing In a context of sustainable public finances, but with to implement fiscal stimulus in crisis times (Cuadra, rising expenditure pressures, the emphasis could be Sanchez, & Sapriza, 2010). Recent evidence shows placed in improving the quality of public spending, that about one-third of developing countries were able as discussed in the rest of this PER. Over the past two to shift to implementing countercyclical fiscal policy after years, efforts to improve spending efficiency continued year 2000, while a few developed countries went “back with implementation of full program-based budgeting, to school” (Frankel, Vegh, & Vuletin, 2013). Cambodia linking budget to policy and strategy. However, budget is one of the countries that has progressed most in fragmentation—with parallel formulation, execution, terms of moving from a highly pro-cyclical fiscal policy and reporting processes for domestically and externally to countercyclical policy in recent years (Figure 16). financed resources—remains one of the main challenges. Some policy options are presented in the As the economy is highly dollarized, fiscal policy next subsection. is one of the main policy tools for macroeconomic management, so maintaining sufficient fiscal space Fiscal policy options going forward is particularly important for Cambodia. In this context, authorities have kept a fixed exchange rate by pegging Cambodia is one of the few developing countries the Cambodian riel to the U.S. dollar to maintain price that has been able to implement countercyclical stability; the exchange rate has been hovering around fiscal policy in recent years. Fiscal policy has been 4,000 riel per dollar for several years. The fiscal multiplier, traditionally a-cyclical or countercyclical in developed measured as the impact on GDP of an additional dollar countries and pro-cyclical in emerging countries, in government spending, has been shown to be zero as discussed in Kaminsky et al. (2004). It has been under a flexible exchange rate and positive under argued that developing economies usually face political fixed exchange rate regimes. The average multiplier of 22 Cambodia: A Public Expenditure Review Figure 16. Correlation between GDP and the structural component of public expenditure 1 Back to school Still in school Greece Venezuela 0.75 Uruguay 0.5 Corr(G, GDP) 2000-2014 0.25 Colombia Brazil Italy Portugal 0 Peru -0.25 Australia India -0.5 UK Cambodia Finland -0.75 Chile USA Established graduates Recent graduates -1 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 Corr(G, GDP) 1960-1999 Source: Carlos Vegh. Update to figures in Frankel, Vegh and Vuletin, 2013. public spending in a developing economy has been estimated to be as high as 50 or 60 percent of the funds calculated in 0.57 in the first quarter after the stimulus invested, levels that would severely hamper any fiscal is applied and 1.6 in the long run (Ilzetzki, Mendoza, & stimulus efforts. The rest of the chapters in this PER aim Végh, 2010). In Cambodia, lack of accurate GDP data at increasing value for money in tax expenditure, PIM, on the demand side prevents the calculation of such and sectoral expenditure allocation and execution. multipliers. Nonetheless, being an economy with low debt levels operating under a fixed exchange rate, there A number of policy options could help Cambodia would in principle be a case for Cambodia to use public further increase the impact of fiscal measures capital spending as a countercyclical tool in moments of going forward, now that authorities are implementing economic slowdown. expansionary fiscal policy for the first time in a few years and in the context of declining fiscal space: Notably, the impact of fiscal measures may not be as large as expected if there are leakages, making First, continued efforts to implement the Public it important to identify and address spending Financial Management Reform program, including inefficiencies. Evidence suggests that the positive a functioning MTBF, can help boost value for stimulus of fiscal policy can be much larger in industrial money. Further improvements need to be pursued at countries than in developing economies (Riera- different stages of the budget cycle. On the planning Crichton et al., 2014). This can be explained by the side, effective adoption of a Medium-Term Fiscal fact that developed economies are often more efficient Framework and a Medium-Term Budget Framework, in allocating and executing spending. Waste due to including expenditure ceilings for line ministries, can inefficiencies in some developing economies has been help making BSPs more realistic. Ideally, BSPs in all line Cambodia: A Public Expenditure Review 23 ministries would integrate DP-funded projects. Other would be to leverage public-private partnerships areas for improvement include budget execution (with (PPPs) to compensate for the eventual decline in DP- full implementation of the FMIS system and streamlined financed capital projects. This requires close attention approval processes); accountability (budget managers to the development and implementation of the legal become responsible for financial management and and regulatory framework pertaining to concessions, results); and monitoring and evaluation (internal control government guarantees, and state property control, and audit, budgetary oversight). A more predictable and as well as debt and contingent liability management. transparent use of public funds could be achieved by Cambodia would also need to lay the foundation for providing more timely budget allocations to executing domestic bond issuances. Overall, regardless of the units and avoiding the practice of keeping books ultimate mix of sources of funding, this transition is open until the middle of the next natural year. All expected to require stronger project selection criteria these elements would be needed to effectively adopt and public debt management. performance-informed budgeting in the medium to long term. In addition, an improved public investment management framework is needed to scale up Second, measures should be put in place to ensure government-financed capital expenditure. Currently, that increases in wages translate into better quality the capacity of most line ministries to implement medium- services to citizens. While the size of the public sector and large-scale investment projects is very limited, and remains modest by international standards, quickly institutional coordination across the different stages rising salaries are likely to result in fiscal pressures. In the of the public investment cycle (planning, budgeting, short term, there is a need to carry out an assessment implementation, monitoring) seems to be almost non- of the allocation of public employees, as well as a review existent (see Chapter 3). Going forward, selection of of the functions and operations currently performed new public investments should be prioritized carefully by different departments and units (including at the and made contingent upon sufficient funding for O&M in subnational level), in order to optimize the use of the coming years. resources and bring them in line with sector strategic objectives and BSPs. Making future wage increases Finally, tax expenditure should be monitored and and incentives conditional on public sector performance revised as a source of potential revenue gains. indicators would help enhance the efficiency of public Implementation of the Revenue Mobilization Strategy spending. This could be coupled with strengthening approved in 2014 has so far resulted in a rapid increase of citizen feedback channels and managers becoming in revenue collection, allowing for continued fiscal accountable for program outcomes. consolidation despite growing expenditure levels. Further progress in tax administration reform under this Third, the changing terms of capital financing framework could potentially be complemented by the will require further strengthening of public debt streamlining of existing tax incentives to keep increasing management. Cambodia’s graduation as a lower revenue collection over the medium term, as discussed middle-income economy in 2015 has already resulted in Chapter 2. in a gradual shift from grant funding to concessional loan financing by DPs. This trend is likely to continue as traditional DPs re-align their financing packages in view of Cambodia’s eventual graduation from the International Development Association, which would naturally bring costlier borrowing terms. An alternative 24 Cambodia: A Public Expenditure Review Table 8. Policy options towards a more effective and sustainable fiscal policy Challenge Short-term policy options Medium-term policy options (1-2 years) (3+ years) The budget remains Effectively implement a MTBF, including Full implementation of performance-in- fragmented and weakly budget ceilings to line ministries for more formed budgeting linked to planning realistic BSPs Quickly rising public Conduct a functional and operational Link wage increases/incentives to perfor- sector wages review in all ministries, and optimize the mance allocation of staff, in line with strategic objectives Decline in concession- Provide capacity building on manage- Develop a domestic debt market and al component of DP ment of contingent liabilities as PPPs issue public bonds financing expand Source: World Bank staff elaboration. Cambodia: A Public Expenditure Review 25 2 BOOSTING REVENUE COLLECTION BY STREAMLINING TAX INCENTIVES Introduction and motivation the public payroll has been rising, from 5 percent of GDP T in 2013 to a forecasted 8-9 percent of GDP by 2018, and the increase in the public sector minimum wage as hanks to a series of tax administration discussed earlier could result in some fiscal pressures. reforms, Cambodia has been able to boost its revenue collection significantly In this context, Cambodia has an opportunity to in recent years. Among other reforms, since 2008, increase revenue collection without raising taxes, the General Department of Customs and Excises by rationalizing tax exemptions. Drawing from import has successfully rolled out the Automated System for transactions data provided by the General Department Custom Data (ASYCUDA) as well as enhanced anti- of Customs and Excises (GDCE), tax expenditure smuggling measures. More recently, in the context of at customs was an estimated 3.9 percent of GDP in the Revenue Mobilization Strategy for 2014-18, the 2015 and is granted mainly in the form of exemptions General Department of Taxation (GDT) emphasized to Qualified Investment Projects (QIP). In addition, a shift toward a real regime and has improved its previous analysis estimated tax expenditure in the form capacity in tax auditing and taxpayer services, leading of Corporate Income Tax (CIT) exemptions to QIPs to better compliance. Thanks to such improvements in the garment sector at 1.8 percent of GDP in 2014 in tax administration and a broadening of the tax base, (World Bank, 2015). The collection of direct taxes— Cambodia’s tax revenue collection increased from 12.7 largely taxes on profits (corporate income tax)—has percent of GDP in 2013 to 16.1 in 2016.9 been increasing but remains relatively small due to the exemptions provided under the existing investment law, Further increases in revenue collection are needed to as well as challenges to compliance. support Cambodia’s development, especially given the rapid increase in payroll pressures. Cambodia This chapter proposes potential ways to streamline faces two medium-term challenges related to revenue existing tax incentives without diminishing the collection: (i) further erosion in the grant component of country’s attractiveness to foreign direct investors. revenue and (ii) an expected decline in trade taxes, in Section 2.2 describes recent revenue collection the context of tariff reduction associated with progress performance and discusses improvements in tax in ASEAN integration and participation in WTO rounds administration, as well as potential reforms going forward. as a middle-income country. On the expenditure side, 9  According to updated official figures provided in 2019. 26 Cambodia: A Public Expenditure Review Section 2.3 analyzes existing incentives in Cambodia, decade) (Figure 17, left panel). As described above, the benchmarking them with respect to other East Asian RGC has responded with efforts to strengthen the tax countries, and estimates tax expenditure based on administration and boost tax revenue. Coupled with fast customs transaction data. The Chapter concludes by economic growth, these efforts have helped increase, presenting a series of potential measures to streamline and Cambodia has thus surpassed the average for existing incentives according to international best lower middle-income economies (Figure 17, right panel). practices. Nonetheless, tax revenue remains behind the levels observed in Lao PDR and Thailand. Both the General Department of Taxation (GDT) and Cambodia: stellar gains in revenue the General Department of Customs and Excises collection (GDCE) have succeeded in improving revenue collection in recent years. Improving taxpayers’ Improvements in tax administration have registration and services, including the introduction boosted collection in recent years of online tax payments and awareness campaigns, has underpinned increased filing of tax returns. Tax To compensate for the decline in foreign aid grants, payments can now be made at commercial banks with Cambodia has noticeably improved tax collection improved services. Stronger tax auditing and arrears over the past five years and is now above the average collection, undertaken by more qualified and motivated for lower middle-income countries. As discussed tax officials and facilitated by an upgraded IT system, earlier, over the past decade, the grant component have contributed to the increases in collection. Average of foreign aid has declined as a percentage of total annual collection growth rates were 13.6 percent for revenue. The ratio of grants to total revenue declined GDCE and 12.5 percent for GDT during 2013-15, and from around 42 percent in 2002 to 25 percent in 2016, collection levels by the GDCE and GDT reached an although it remains above the average for lower middle- estimated 8.6 percent of GDP and 6.5 percent of GDP income economies (around 20 percent over the past in 2016, respectively (Figure 18). Figure 17. Faced with declining foreign aid grants, Cambodia has boosted its tax revenues Grants & non-tax to total revenue Tax revenue to GDP 45 16 40 14 35 12 Percent of Revenue 30 Percent of GDP 10 25 8 20 6 15 4 10 5 2 0 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Cambodia KHM Lower middle-income Cambodia KHM Lower middle-income Source: World Development Indicators, TOFE for 2015 and 2016. Cambodia: A Public Expenditure Review 27 Figure 18. Both inland and customs revenue Figure 19. Tax collection has improved have significantly increased in real terms across different types of taxes Tax revenue as % of GDP, by administration Evolution of tax revenue as % of GDP 10 16% 14% 8 12% 6 10% 8% 4 6% 4% 2 2% 0 0% 2009 2010 2011 2012 2013 2014 2015 2016 2009 2010 2011 2012 2013 2014 2015 2016 Customs Department Tax Department Direct tax revenue VAT Excise Other indirect Tax on trade Source: World Bank staff calculations based on TOFE (Ministry of Source: TOFE (Ministry of Economy and Finance), NIS for GDP Economy and Finance). figures. Thanks to these efforts, collection has increased remained the largest source of government revenue, across different types of taxes. Gains have been at around 35 percent of the total (Figure 20, left realized for most sources of revenue, particularly VAT panel). The share of other domestic taxes (on specific (collections increased from 3.7 to 5.1 percent of GDP goods and services12) also remained stable at 21.5 between 2011 and 2016) and corporate income tax percent, increasing to 24 percent in 2015 and (from 1.4 to 2.7 percent of GDP; Figure 19). Trade 2016. Over this same period, direct taxes gained tax collection has remained around 2.5 percent of share in total collections, with corporate income tax GDP despite a progressive decline in tariffs aimed collections increasing from 12.8 to 17.9 percent and at meeting ASEAN and WTO agreements, which is payroll tax increasing from 4.2 to 4.6 percent of total expected to continue into 2018.10 In fact, thanks to revenue. Although direct tax collection, at 3.6 percent numerous factors, 11 revenue collected by the customs of GDP, remains relatively low in comparison to other administration has continued to expand in both nominal countries in the region (e.g., 5.6 percent of GDP and real terms even with the decline in tariffs. in Indonesia and 7.4 percent of GDP in Thailand), the rapid improvement indicates good progress in The overall structure of domestic revenue has administration. Notably, the share of trade-related tax been shifting, with a rising share of direct taxes declined over the same period (Figure 20, right panel), and a declining share of trade taxes. Looking at although collections remained stable as a percentage the 2010-2016 period, value added tax collection of GDP. The authorities’ efforts to gradually transform 10  According to the General Department of Customs and Excises, and under preferential treatment, Cambodia shall implement the ASEAN Trade in Goods Agreement (ATIGA) by 2015. In this context, Cambodia has implemented tariff reduction since 2009 and shall eliminate import duties on all products by 2015, with flexibility to 2018 (7 percent of the total tariff lines or 662 tariff lines will remain at a 5 percent rate until 2018). 11  The tariff reductions have been partly offset by implementation of administered prices (instead of transaction prices) in the collection of various customs duties and taxes on key import items such as vehicles and petroleum products. An additional tax on oil, intended to be earmarked to road maintenance, is also charged along with the duties that are specific to oil products. Moreover, additional customs duties of 20 percent or higher apply to luxury vehicles. Substantial improvements in the collection of VAT and excise taxes on imports, together with the rollout of the ASYCUDA system, have also contributed to improved collection of customs duties and other taxes on imports. 12  According to authorities, these taxes are imposed on goods that produce negative externalities. This includes oil products as well as non-oil products such as beer, cigarettes, and alcoholic beverages. 28 Cambodia: A Public Expenditure Review Figure 20. The overall structure of tax revenue has been shifting, with direct taxes gaining share while the share of trade taxes has fallen Struture of tax revenue, % of total Trade and direct taxes, as % of total 100% 30% 25% 80% 20% 60% 15% 40% 10% 20% 5% 0% 0% 2009 2010 2011 2012 2013 2014 2015 2016 2009 2010 2011 2012 2013 2014 2015 2016 Payroll Corp inc VAT Tax on imp Trade tax total Direct tax revenue Tax on exp Excise Other indirect Source: World Bank staff calculations based on TOFE (Ministry of Economy and Finance). Table 9. Cambodia’s tax administration processes are still viewed as relatively burdensome Doing Business sub-index rankings and indicators for ASEAN countriesa Country Global Ranking Payments Time (hours Total Tax Rate Post-filing on Ease of (number per year) (% of profit) Indexb (0-100) Paying Taxes per Year) Singapore 7 5 64 20.3 71.97 Thailand 67 21 262 28.7 73.41 Malaysia 73 8 188 39.2 52.65 Vietnam 86 14 498 38.1 95.71 Philippines 105 20 182 42.9 50 Indonesia 114 43 207.5 30 68.82 Myanmar 125 31 282 31.2 45.54 Cambodia 136 40 173 21.7 25.97 Lao PDR 156 35 362 26.2 18.57 Source: Doing Business 2018, the World Bank Group. Note: a. Except Brunei, as it is not included in the survey. b. Composite measure of the time to comply with VAT or GST refund, time to obtain VAT or GST refund, time to comply with CIT audit, and time to complete CIT audit. The index is based on the distance to the country with the most burdensome procedures. Thus, a higher index implies less time-consuming processes. the previous tax structure, which was heavily sustain collection as the country pursues free trade dependent on trade taxes, to a more advanced tax policy objectives such as those envisaged under the system based on direct and domestic taxes will help ASEAN Economic Community. Cambodia: A Public Expenditure Review 29 However, despite improvements, tax collection same time, other countries with low levels of staffing are and returns procedures in Cambodia are still able to reach a significantly higher number of taxpayers. daunting. Despite having low total tax rates on profits The average number of active taxpayers to staff (Table 9), Cambodia ranks in the lower one-third of among similar countries is 676, whereas in Cambodia countries in the Paying Taxes sub-index of the Doing it is only 28. This is likely to be related to widespread Business report (136 out of 189) due to burdensome informality as well as the lack of a personal income tax. administrative processes. For example, the time According to the World Bank Group Enterprise Surveys, required to comply with and obtain a VAT refund 13 in the percentage of firms in Cambodia declaring to face Cambodia is 21 hours and 63.9 weeks, respectively, informal competitors (nearly 80 percent in 2016) is the compared to 18 hours and 31 weeks in Indonesia. largest in ASEAN. Moreover, it takes 31 hours to comply with, and 39 weeks on average to complete a corporate income Due to the nature of its payroll tax, Cambodia tax audit, substantially longer than in all other ASEAN collects less from individuals than other countries economies. applying personal income taxes. For a nominal rate of 5 percent, payroll tax collections represented around Exemptions continue to limit potential 1.8 percent of GDP in 2012, which is below international collections from existing taxes levels (Figure 21 above, right panel). However, a payroll tax cannot be directly compared to personal income Under the current tax administration structure, taxes, since the latter has a much broader scope, Cambodia could still significantly increase the levying taxes on various sources of income and not only number of active taxpayers. According to the IMF from paid employment. Payroll taxes do not capture Collecting Taxes Database, Cambodia’s tax staff per personal income from unincorporated income, which thousand inhabitants, 0.10 in 2012-13, was well below could be a significant source of revenue collection for a the average country, at 0.65 (Figure 21, left panel). At the country like Cambodia. Figure 21. Cambodia has fewer tax administration staff relative to its population and fewer active taxpayers relative to tax staff than other countries Tax administration staff, 2012-13 PIT rate and collections, 2012-13 2,000 25 1,800 1,600 20 PIT collections in % of GDP 1,400 Active taxpayers to staff 1,200 15 1,000 800 10 600 400 5 200 KHM KHM 0 0 0.0 0.5 1.0 1.5 2.0 0 10 20 30 40 Tax staff per 1000 people PIT rate Source: USAID Collecting Taxes Database. 13  The time required to comply with VAT refund includes the audit process if it is considered highly likely, such as in the case of Cambodia. 30 Cambodia: A Public Expenditure Review Figure 22. VAT and CIT productivity VAT and CIT productivity, 2012-13 1.0 0.9 0.8 0.7 (CIT/GDP)/(CIT rate) 0.6 0.5 0.4 0.3 0.2 KMH 2015 0.1 KMH 2012 0.0 0 50 100 150 200 250 300 (VAT/GDP)/(VAT rate*C) Source: USAID collecting taxes database. Despite recent improvements, the productivity Streamlining tax incentives in of the VAT and CIT in Cambodia remains low by international standards. VAT productivity in Cambodia, Cambodia measuring the effectiveness of tax collections, Existing incentives in Cambodia are increased from 44.2 percent in 2012-13 to an estimated predominantly tax exemptions outside 48.1 in 2015 (Figure 22), while remaining below the the tax code sample average (66 percent).14 Corporate income tax productivity, measured as collections to GDP divided by Investment incentives were first introduced in the statutory tax rate, improved from 0.07 (one a 0 to 1 Cambodia with the promulgation of the Law on scale) in 2012-13 to 0.11 in 2015, although it is still less Investment (LoI) in 1994,15 in the wake of the country’s than half of the cross-country average (0.15). In both first post-war free general election monitored by the UN. cases, existing tax exemptions limit the tax base and the The 1994 LoI stipulated which sectors and geographical scope for gains in tax collections. The following section areas could apply for incentives.16 It granted tax holidays looks more closely at the prevalence of tax incentives of up to 8 years,17 a zero CIT rate for reinvested profits, in Cambodia, which may be partly responsible for the and an exemption from payment of import duties and relatively low effective tax collection relative to nominal VAT on inputs for firms qualifying for tax holidays. rates. 14  The productivity of the VAT in Cambodia is measured by collections to GDP divided by the applied tax rate multiplied by total private consump- tion in the economy (as per the national accounts). 15  Cambodia Law on Investment No 030324, 1994. 16  These sectors and areas included pioneer or high-tech projects, job-creating and export-oriented activities, tourism, agro and processing, infrastructure, energy, rural development, environment, and Special Economic Zones. 17  Cambodia Investment Law: article 13. Cambodia: A Public Expenditure Review 31 Table 10. Most investment incentives in Cambodia have tax implications No. List of incentives available in Cambodia 1 Corporate Income Tax (CIT) exemption 2 Special depreciation 3 Duty-Free Import for Domestically oriented QIP 4 Duty-Free Import for Export oriented QID 5 Duty-Free Import for Supporting QIP 6 Exemption of Export Tax 7 Exemption of Value Added Tax (VAT) for SEZ investors 8 Land Concession for SEZ developers 9 Special Customs Procedures 10 CIT exemption for SEZ developer 11 Import duty exemption for SEZ developers 12 Investment guarantees for SEZ investors and developers 13 VAT exemption for imported production inputs 14 VAT exemptions for garment imported inputs 15 Import duty reducton/exemptions 16 VAT exemption on various agricultural materials such as seeds, breeds, etc. 17 Economic Land Concessions (ELCs) Source: World Bank (2015). As the country continued to develop, the 1994 payments on both their inputs and sales. In addition, LoI was amended in 2003 with the goal of making the 2003 amendment introduced more detailed it more consistent with the Law on Taxation, requirements and procedures to be fulfilled when rationalizing tax incentives, and clarifying criteria applying for tax incentives, including sector-specific and procedures to qualify for incentives. As a result minimum thresholds for investments. of the changes introduced in the 2003 amendment and its implementing sub-decree18 (issued in 2005), Special Economic Zones in Cambodia also benefit the length of tax holidays ranges between 6 to 9 years from specific incentives. A 2005 sub-decree on starting from the first year of sales , depending on 19 Special Economic Zones (SEZs) recognized exemptions the government assessment. The 2003 amendment on import duties, VAT, and CIT for SEZs.20 The sub- also introduced the possibility for investors to choose decree also granted guarantees and land concessions a special accelerated depreciation scheme instead of to investors. The number of SEZs is set to expand, with the tax holiday. Supporting industries to the garment thirty zones approved and nine operational at the end and footwear sectors are also exempted from VAT of 2015. The largest SEZ is the Sihanoukville Special 18  Sub-decree on the implementation of the law on the amendment to the law on investment of the Kingdom of Cambodia No 111 ANK/BK dated September 27, 2005. 19  Trigger period (starting from receiving QIPs status, and lasting until the first profit is made or 3rd year after first years of sales) + 3 year-holiday periods + priority period (1 to 3 years depending on the capital invested). 20  Sub-decree 148 ANK/BK on the Establishment and Management of the Special Economic Zones (SEZs). 32 Cambodia: A Public Expenditure Review Economic Zone, set up in 2008, which housed 93 Incentives are usually granted within a month, and factories in 2017, but with plans for large expansion. most companies do not experience significant delays. In Cambodia, the Investment Law stipulates Overall, the majority of existing investment that the Conditional Registration Certificate (CRC) shall incentives in Cambodia are tax incentives with be issued to the investor within three working days of revenue collection implications, as opposed to the investor’s submission of an investment proposal financial or other types of incentives. According to a to the CDC, the institution in charge of assessing review of investment incentives (World Bank, 2015), the applications to become a Qualified Investment Project majority are found to be incentives with tax implications, (QIP) that is eligible for tax exemptions. The investor is including corporate income tax exemptions, import duty then expected to obtain all necessary licenses from the exemptions, special depreciation, VAT exemption, and relevant ministries/entities listed in the CRC within 28 export duty exemptions (Table 10). Only four of them working days after issuance of a CRC. The analysis of seem to be financial incentives in the form of investment Investor Motivation Survey (IMS) conducted by the World guarantees, special customs procedures, and land Bank in 2015, confirmed that 75 percent of respondents concessions.21 Except for those incentives offered did not report a delay in receiving approval of their QIP to investors in SEZs, existing schemes are generally application. The remaining 25 percent of firms did report economy-wide and offered at the national level.22 Yet a delay of 14 days on average. The percentage of firms exemptions have not been reflected in the tax code, reporting delays did not vary significantly across sectors as per international best practices. The provision of or firm size except for firms with over 2,000 employees, widespread incentives is not unusual in East Asia and of which only 18 percent reported delays (World Bank, the Pacific, but they may not be the most effective 2015). instrument to promote foreign direct investment attraction (Box 2). Nonetheless, most surveyed companies incurred some costs to process their QIP applications, Tax incentives in many ASEAN countries are which could potentially pose a barrier to entry for designed to generate public goods and externalities, smaller firms. Most of the companies hired a broker while in Cambodia seem to be at the moment more to process their applications, while a few reported supportive of employment generation. Such positive additional costs. Around 72 percent of the surveyed externalities include the development of remote regions, firms reported that they had hired a broker or other priority or high-technology industries, and other public third party to process their QIP applications. It is goods. Tax incentives in Cambodia do not yet target estimated that it costs about USD 12,000 for firms such additional objectives. Garment and footwear in the garment and footwear industry and about USD projects represent the majority of QIP (383 out of 716 20,000 for those in the other manufacturing sector QIPs in 2010-14), but their share of total registered fixed to process their investment incentives package. assets was just 11 percent, indicating that incentives This could potentially pose a barrier to entry of are mainly directed to a non-capital-intense industry. SMEs. Roughly one-fifth reported at least one minor Nonetheless, the creation of employment is often additional cost, and around five to ten percent greater than in other sectors, and the cost per job of reported at least one major cost associated with the QIP incentives has been estimated in less than USD 500 application process (additional consulting fees, losses per year (World Bank, 2015). associated to delays, etc.). 21  Desk research and project interviews to date have not identified any financial incentives (such as matching grants, subsidies, etc.) that might be granted to firms by individual ministries and government entities, other than the incentives listed on the CDC website (and mentioned above). 22  According to the CDC investment guidebook of 2013, certain investment projects are not eligible for QIP incentives. These include imports and exports for commercial purposes, wholesale and retail, transportation, tourism, financial business, professional services, hotels below 3-star grade, and real estate development. Cambodia: A Public Expenditure Review 33 Box 2 ASEAN’s “race to the bottom” in investment incentives Around the globe, countries grant incentives with the expectation of attracting foreign investment, which can help increase the size and quality of the capital stock and lead to positive externalities in the form of backward linkages to local industries and knowledge spillovers. Developing nations compete to attract FDI both through low wages, low labor, and environmental standards and through tax breaks (Chau & Kanbur, 2006; Hecock & Jepsen, 2013; Klemm & Van Parys, 2012). This has led to a “race to the bottom,” in which countries make themselves worse off collectively. In Africa, effective tax rates have effectively fallen to zero in industries where special regimes are in place (Abbas & Klemm, 2013). In Barbados, in the face of tax competition and aggressive bargaining by multinational hotel chains, offers tax holidays of up to 40 years for hotel investments. The most common incentives used by countries in ASEAN are tax holidays, reduced tax rates, and research and development (R&D) incentives. In addition, compared to other regions, East Asia and the Pacific has the highest share of countries granting discretionary incentive treatment, and incentives are usually provided outside of the tax code. ASEAN economies offer five to eight types of tax incentives to firms on average. In contrast, advanced East Asian economies such as Japan and Hong Kong SAR offer only two or three types of incentives (and no tax holidays). As a result, tax collection in most economies in Southeast Asia is below 20 percent of GDP, which calls for talks in ASEAN leading to harmonization of investment incentives to level the playing field and reduce tax competition (World Bank, 2015). Cambodia is very much in line with the regional trend of offering a large number of incentives to investors. These include tax holidays (which, at up to 9 years, are on the generous side), reduced rates, tax credit, specific incentives for SEZs, and other discretionary incentives. Notably, Cambodia—together with Indonesia and Myanmar—is one of the few developing countries surveyed that does not currently offer incentives for R&D. Tax holidays in Cambodia could potentially be unnecessary for many investors: approximately 75 percent of the firms surveyed in the garment, footwear, and other manufacturing sectors did not consider any alternative location to Cambodia in their investment decisions (World Bank, 2015). The international experience on the effectiveness of incentives to attract FDI is mixed at best, and tax breaks cannot compensate for challenges in investment climate. World Bank Investor Motivation Surveys in more than 15 economies find that, overall, tax incentives are not the main attraction factor for surveyed investors. Tax incentives seem to contribute to FDI attraction in advanced economies, but the response is limited in developing economies (Klemm & Van Parys, 2012). Empirical evidence also shows that a weak investment climate or lack of infrastructure hampers the efficacy of tax incentives (James & Van Parys, 2009). Finally, tax incentives are found to be mostly effective in attracting efficiency-seeking FDI such as, for example, garment producers that are looking to minimize costs and export to developed countries, while they seem to be generally ineffective in attracting market-seeking or resource-seeking FDI (Grubert & Mutti, 2004). 34 Cambodia: A Public Expenditure Review What is the cost of investment incentives 2015 and is above actual tax collections. This is in Cambodia? higher than a previous World Bank (2015b) estimate of tax relief at customs (4.1 percent of GDP in 2014), Even when incentives are justified and effective in since this analysis looks at relief granted to other actors terms of attracting FDI, they carry an opportunity aside from QIP (including SEZs, NGOs, and others) and cost since they result in lower revenue collection. comprises a more detailed analysis of the different relief When tax exemptions are granted to a group of investors, categories, prepared in collaboration with GDCE. Actual the tax burden on the rest of the taxpayers increases, if tax collections (excluding excises) at customs expanded the level of government spending remains constant. For from 4.6 percent of GDP in 2014 to 5.3 percent of GDP this reason, many countries opt to quantify the implicit in 2015 (Figure 23). tax expenditure and publish annual statements in the national budget, thus giving a sense of how much public The bulk of tax revenue relief is explained by funding is implicitly subsidizing the beneficiary industries exemptions on VAT and customs duties. Tax on value and increasing transparency and awareness. added represented 41 percent of total collections in 2014 and 2015, followed by custom duties (30 percent) This subsection assesses the cost of investment and special duties (28 percent; Figure 24, left panel). At incentives in Cambodia. The analysis presented the same time, most of the tax relief is granted on value provides partial estimates of tax expenditure in added tax (52 percent of total relief) and customs duties Cambodia and is based on customs transaction data (42 percent), whereas special duties only comprise 5.6 from the ASYCUDA system for 2014 and 2015 provided percent of the total. While the ratio of tax relief almost by the GDCE. The customs administration collects six doubles collections in the cases of customs duties different taxes at the border: customs duties, special and VAT, it only represents 26 percent of special tax duties, VAT, and three petroleum-specific taxes. 23 collections, for which exemptions are less prevalent The analysis presented in this subsection focuses on (Figure 24, right panel). Petroleum excise relief is minimal the first three taxes. Consistency issues in currently available tax return information preclude the elaboration Figure 23. At customs, tax relief increased of comprehensive tax expenditure estimates on VAT faster than collections in 2015 (beyond customs) and CIT. Tax collected at customs vs. tax relief at customs The analysis aims to calculate tax relief at customs, 8% 7.6 which constitutes the implicit government expenditure that goes through the tax system. 6% 5.5 5.3 Cambodia classifies imports into more than 400 % of GDP 4.6 categories depending on the type of relief offered. For 4% the purposes of this paper, these relief categories are grouped according to beneficiary: relief for Qualified 2% Investment Projects, relief for firms in Special Economic Zones, and relief for other organizations (including government, DPs, NGOs). 0% 2014 2015 Tax collected Total tax relief Foregone revenue at customs increased from 5.5 Source: World Bank staff calculations based on data from the General percent of GDP in 2014 to 7.6 percent of GDP in Department of Customs and Excises. 23  The VAT in Cambodia applies a general rate of 10 percent. Custom duty rates are 0, 7, 15, or 35 percent. Special taxes applied on top of custom duties vary from 5 to 50 percent, depending on the product. Cambodia: A Public Expenditure Review 35 Figure 24. Customs duties and VAT account for most of the tax relief in Cambodia Breakdown of revenue collected and relief, 2014+2015 Ratio of relief to collection, by tax 10,000 200% 180% 8,000 160% 140% Billions of Khmer riel 6,000 120% 100% 4,000 80% 60% 2,000 40% 20% 0 0% Tax collected Tax relief Customs Duties Special VAT Customs Duties Special VAT Source: World Bank staff calculations based on data from the General Department of Customs and Excises. (0.02 percent of GDP in 2015, compared to collections refunds promptly, this advantage would be minimal. over 1 percent of GDP), with aid and DP-related projects Only VAT relief on imports by final consumers should being the main beneficiaries. be considered a tax expenditure, with two types of final consumers: (i) both households purchasing goods or QIPs are the main beneficiaries of tax incentives in services for their own consumption, which should be Cambodia. QIPs were beneficiaries of KHR 4.6 billion taxed, and, on the other hand, (ii) businesses that are in tax relief in 2015, or 6.3 percent of GDP (Figure 25). exempt for achieving a certain purpose. For example, By tax, QIPs accounted for around 86 percent of VAT in Cambodia, “domestic public goods” like health care and 83 percent of customs duties relief. In the case of and education are exempt. In most countries, financial relief on special duties, SEZs and other actors (including NGO, DPs, and government) hold a larger share, with Figure 25. QIPs and firms in SEZs are 34 percent and 5 percent, respectively. the main beneficiaries of tax incentives VAT tax relief on imports is not usually considered an Tax relief, by beneficiary, 2014+15 actual tax expenditure when the goods acquired are 6,000 inputs to the production process. The VAT is among Billions of Khmer riel 5,000 the most efficient taxes because it aims to charge the final consumer, avoiding tax cascading by eliminating 4,000 the tax liability on production, lowering production costs, 3,000 and reducing disincentives to productivity-enhancing investments on capital, machinery, and equipment. Tax 2,000 relief on VAT for imports is not usually granted, since the 1,000 input tax credit mechanism ensures that importers can claim the amount back when filing its VAT supported and 0 Customs Duties Special VAT charged to the tax authority. There may be a cash flow Others (NGO, donor, gvt.) QIP SEZ and other firm relief advantage to relieving the VAT at imports, but if firms can Source: World Bank staff calculations based on data from the General claim their input tax credits quarterly and the state pays Department of Customs and Excises. 36 Cambodia: A Public Expenditure Review Figure 26. Tax relief on special and customs duties to SEZs noticeably increased in 2015 Tax relief on special duties, by beneficiary Tax relief on customs duties, by beneficiary 1.0% 4% 0.8% 3% Percent of GDP Percent of GDP 0.6% 2% 0.4% 1% 0.2% 0.0% 0% 2014 2015 2014 2015 Other including govt. QIP SEZ and other firm relief Other including govt. QIP SEZ and other firm relief Source: World Bank staff calculations based on data from the General Department of Customs and Excises. services are exempt because it is usually difficult to (Figure 28). QIP imports of energy generation-related determine the value added on these commodities. In machinery ranked second in tax relief (KHR 225 billion), either case, the value added by the firm making the since Cambodia is at the moment undertaking large exempt supply is not subject to the VAT, and, as a result, projects aimed at increasing energy generation capacity. the tax paid by purchasers of the supply is reduced. Tax relief on the acquisition of electrical machinery and equipment (126 billion) ranked fifth. These results are Tax relief on special duties and customs duties has not surprising, given that manufacturing activity in been on the rise, due in part to an increasing number Cambodia is dominated by the garment sector. Overall, of firms benefiting from SEZ-related exemptions. and even when there is not available information on Unlike in the case of VAT, tax relief on special taxes the sector of operation of the importing firms, the and customs duties is considered a tax expenditure. nature of the imported goods suggests that QIP relief Tax expenditure on special duties jumped from about is supporting the acquisition of intermediate inputs and 0.1 percent of GDP in 2014 to nearly 0.6 percent of machinery for the production process, which would be GDP in 2015 (Figure 26, left panel). Tax expenditure on consistent with one of the purposes of the QIP policy: to customs duties increased from 2.3 percent of GDP in reduce the cost of inputs for companies participating in 2014 to 3.1 percent of GDP in 2015, with QIPs being the global value chains. Nonetheless, further analysis using main beneficiaries (Figure 26, right panel). Notably, tax customs data to trace imported goods by exporting expenditure by firms in SEZs was negligible in 2014 then companies is needed to confirm this assessment. expanded to a combined 0.7 percent of GDP in 2015. Meanwhile, tax relief to other beneficiaries (SEZs, Overwhelmingly, QIP tax relief is granted on imported DPs, government) is mostly granted for the import inputs for the garment industry coming from China of vehicles and energy generation machinery. When (Figure 27). Of the 14 top imports receiving QIP relief looking at SEZs (some of them QIPs) and other non-QIP at customs, 8 of them seem to be related to garment exemptions (DPs, government), the bulk of imported production (e.g., fabrics, fibers, cotton, fur, skins). Tax items are vehicles, energy generation machinery, and relief to the top QIP-imported commodity, knitted or articles of plastic (Figure 29, left panel). Tax relief on crocheted fabrics, amounted to KHR 561 billion in 2015 automobiles totaled KHR 280 billion in 2014 and 2015 Cambodia: A Public Expenditure Review 37 Figure 27. China (including Hong Kong) is the origin of two-thirds of QIP relief import value QIP relief, by country of origin of imports, 2015 China Hong Kong Taiwan Vietnam Thailand, Kingdom of Korea, Republic of Malaysia Japan Singapore USA India Germany 0 200 400 600 800 1,00 1,200 Source: World Bank staff calculations based on data from the General Department of Customs and Excises. Figure 28. Garment imports dominate QIP relief at customs QIP tax relief, top imported commodities, 2015 Knitted or crocheted fabrics Nuclear reactors, boilers, machinery and ... Man-made staple fibres Plastics and articles thereof Electrical machinery and equipment and ... Cotton Furskins and artificial fur; manufactures thereof Paper and paperboard; articles of paper pulp, of paper ... Special woven fabrics; tufted textile fabrics;... Articles of apparel and clothing accessories,... Rubber and articles thereof Miscellaneous manufactured articles Raw hides and skins (other than furskins) and leather Impregnated, coated, covered or laminated... 0 100 200 300 400 500 600 Billion of Khmer riel Source: World Bank staff calculations based on data from the General Department of Customs and Excises. combined. This relates to existing exemptions on In terms of CIT, Cambodia’s statutory CIT rate of vehicle imports for government, DPs, and NGOs as 20 percent is among the lowest in the Asia-Pacific well as by some individuals (Figure 29, right panel). It 24 Region. In fact, it is lower than CIT rates in other ASEAN would be worth conducting a detailed analysis of the economies except Singapore (17 percent) and Thailand vehicles imported, since these kinds of exemptions (also 20 percent). The country’s effective average tax often result in loopholes of which both individuals and rate (EATR), a measure of the tax burden of investments firms take advantage to avoid paying taxes on activities which incorporates other relevant features of a country’s that should be taxable. 24  By looking for identifier 999999 in customs data, it is possible to estimate tax expenditure on vehicles imported by individuals, amounting to an estimated 0.33 percent of GDP in 2015. Over half of the relief relates to exemptions on special taxes (Figure 29, right panel). Given that households are not eligible for VAT input tax credits (since they import for consumption purposes), all the relief granted to them is considered to be a true tax expenditure. For the purposes of this analysis, VAT relief on individuals (identifier 999999) is considered a tax expenditure (estimated in 0.09 percent of GDP in 2014 and 0.15 percent of GDP in 2015). 38 Cambodia: A Public Expenditure Review Figure 29. Tax relief on imports of vehicles rank first on non-QIP exemptions Tax relief for SEZs and to other non-QIP, Tax expenditure on imports top imported commodities, 2014 and 2015 by individuals, 2015 0.35 Vehicles other than railway or... Energy generation machinery 0.30 Plastics and articles thereof Nuclear reactors, boilers,... 0.25 Miscellaneous manufactured articles % of GDP 0.20 Man-made staple fibres Knitted or crocheted fabrics 0.15 Cotton Rubber and articles thereof 0.10 Articles of iron or steel Furniture; bedding, mattresses,... 0.5 Paper and paperboard; articles of... 0.0 0 100 200 300 Rest Automobiles Billion of Khmer riel Customs Duties Special VAT Customs Duties Special VAT Source: World Bank staff calculations based on data from the GDCE. tax code beyond CIT rates (such as depreciation and such as Germany or Singapore, where dividend payments other capital allowances, interest deductibility, and received by parent companies are exempt from taxation, property taxes), is also low compared to other countries Cambodia’s EATR is higher than in Myanmar, Vietnam, in the region.25 Using standard assumptions of firm and Malaysia, as these countries do not levy withholding composition and profitability, Cambodia’s EATR is taxes on repatriated dividends (Figure 31). estimated to be 18.4 percent, even lower than the CIT rate. This is the fourth lowest among the Asia-Pacific Tax incentives in Cambodia significantly reduce the economies considered, higher only than Hong Kong, EATR for eligible QIPs, but no more than in other Taiwan, and Thailand (Figure 30). 26 countries in the region. For example, a tax holiday of six years reduces the EATR from 18.4 percent to Cambodia’s CIT effective average tax rate is higher 12.2 percent.28 This is slightly higher than EATRs for for foreign investment repatriating net income. investments enjoying the most generous tax incentives A 14 percent withholding tax rate for dividends and in other countries in the region, including Vietnam, interest paid to parent companies abroad leads to an Thailand, and Lao PDR (9.2 percent, 9.6 percent, and EATR of approximately 30 percent (with small variations 10.6 percent, respectively), but lower than in Malaysia depending on the country of origin) for investments (15.6 percent) and the Philippines (22 percent) (Figure 32). seeking to repatriate net income. This is similar to the 27 EATR in other ASEAN countries for companies based in In sum, at customs alone, tax expenditure in the United States, where foreign affiliates are not exempt Cambodia in 2015 would have amounted to around from taxation at home but can only deduct taxes paid 3.9 percent of GDP, with the bulk of exemptions abroad. For investments originating in other countries benefiting QIP investors (mostly in the garment 25  The computation of the EATR is based on significant features of the respective country’s tax system (e.g., depreciation allowances, profit and non-profit tax rates, exemptions). For full details on the estimation method, see Wiedemann and Finke (2015). 26  While having a lower statutory CIT rate, Singapore’s EATR is higher than Cambodia’s given that the depreciation of buildings and acquired tangibles are not tax-deductible in Singapore. 27  A 10 percent withholding tax rate is applied to four countries with which Cambodia has tax agreements (Brunei, China, Singapore, and Thailand). 28  For firms not repatriating income. Cambodia: A Public Expenditure Review 39 industry). Leaving aside VAT-related relief at customs tax expenditure, data quality issues prevented a full (except those granted to individuals/final consumers) assessment; previous analysis by World Bank (2015) which would be otherwise expected to be claimed estimated CIT tax expenditure by QIP projects in the as input tax credit by importing companies, tax garment and footwear sector at around 1.8 percent of expenditure at customs is estimated to have increased GDP in 2014. Going forward, authorities could consider from around 2.6 percent of GDP in 2014 to 3.9 percent further streamlining existing incentives to minimize the of GDP in 2015 (Table 11). This is larger than the entire fiscal cost and to maximize the potential for attracting budget to the education. With regard to CIT-related higher value-added industries. Figure 30. Cambodia’s statutory CIT rates and estimated EATR are low compared to other Asia-Pacific economies 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% Hong Kong Taiwan Thailand Cambodia Singapore Russia Vietnam Laos South Korea Malaysia China Indonesia Mongolia New Zealand Australia Philippines Japan India Myanmar Effective Average Tax Rate Combined Statutory Tax Rate Source: Wiedemann and Finke (2015). Figure 31. EATR for investments repatriating income is in turn higher than in other ASEAN countriesa 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% Malaysia Laos Vietnam Thailand Cambodia Myanmar Indonesia Singapore Philippines EATR for US Parent Company EATR for German Parent Company EATR for Singaporean Parent Company Source: Wiedemann and Finke (2015). Note: a. Excluding Brunei Darussalam, as it was not included in the study. 40 Cambodia: A Public Expenditure Review Figure 32. When taking into account tax incentives, the EATR comes down to 12.2 percenta 35% 30% 25% 20% 15% 10% 5% 0% Hong Kong Thailand Cambodia Singapore Russia Vietnam Laos South Korea Malaysia China Indonesia Mongolia Philippines Myanmar EATR with Incentive Resulting in Lowest Tax Burden EATR without Incentive Source: Wiedemann and Finke (2015). Note: a. Except Brunei Darussalam, as it was not included in the study. Table 11. Tax expenditure at customs is estimated at around 3.9 percent of GDP in 2015 In percentage of GDP 2014 2015 VAT relief at customs 3.07 3.96 o/w VAT tax expenditure going to final consumers (individuals) (i) 0.09 0.15 Duties and specials, relief at customs (ii) 2.48 3.72 Exemptions on excises (iii) 0.03 0.02 Estimated tax expenditure at customs (i+ii+iii) 2.60 3.89 Estimated CIT tax expenditure for garment and footwear 1.8 1.8a CIT Exemptions for Other Investment ? (Unknown) ? (Unknown) VAT Exemptions for Non-Exporters ? (Unknown) ? (Unknown) Memo item: estimated QIP investment approvals 1.06 Memo item: tax expenditure on the acquisition of automobiles by 0.06 0.33 individuals Source: World Bank staff calculations based on data from GDCE, and World Bank (2015). Note: a. Assuming no variation in 2015 with respect to 2014. Cambodia: A Public Expenditure Review 41 Potential reforms for streamlining activities unlikely to generate wider economic and social investment incentives in Cambodia benefits (James, 2014). As Cambodia tries to move up the value chain and diversify its exports base, improving the investment climate and minimizing the cost of firm Introducing cost-effective incentives and operation will remain priority reforms. This is explicitly rationalizing tax holidays acknowledged in the Industrial Development Policy, which prioritizes the need to ”further strengthen favorable Going forward, Cambodia may face increasing environment for investment and doing business by challenges in retaining the footloose garment improving the regulatory framework, rationalizing the industry, while it is expected to attract higher provision of incentives for investment projects” (RGC, value-added investment. Preferential trade treatment 2015b). under the US quota system, first, and the Everything But Arms agreement with the EU, later, has been the In this context, Cambodia could consider rethinking main factor behind the emergence of the garment its tax incentives mix, in order to limit revenue industry in Cambodia (Bargawi, 2005; Hill & Menon, foregone while continuing to attract investment 2014). In a context of rising wages and relatively high according to a single policy objective. In line costs to firm operation (including electricity, exporting), with the IDP, this policy objective could be to foster the labor-intense low-end garment industry has little capital investment and technology adoption. Evidence incentive to undertake additional investments and suggests that gross capital formation as well as capital scale up production processes, and may be eventually intensity in manufacturing—which are important for relocating. Cambodia, on the other hand, can diversify productivity growth—have been lower in Cambodia and create new clusters around a series of emerging compared to other peer countries (World Bank, 2017). export products that have been growing rapidly Going forward, as Cambodia tries to diversify its in recent years—such as bicycles, ignition wires, manufacturing base, introducing incentives for capital optical appliance parts, electric motors (including for acquisition would be more effective than tax holidays in smartphones). It is nonetheless worth mentioning that fostering longer-term and higher value-added investors the production processes taking place in Cambodia (World Bank, 2015). mainly involve manual assembly, with limited mechanization. This suggests the need for capital Given that they are not linked to policy objectives, investments to improve productivity and value-added, CIT holidays could be progressively replaced or leading to higher wages and salaries. limited to a reduced number of years. For each of the four investors surveyed in the IMS, three were not As incentives are only one of the many factors considering alternative locations to Cambodia. Ideally, that encourage investment, they should be used tax holidays would be targeted to the marginal investor to correct specific market failures (e.g., fostering who would not have invested in Cambodia without investment in machinery, skills training). Currently, incentives. However, lack of sufficient information on incentives in Cambodia are not being used to achieve potential investors (not yet established in Cambodia) targeted policy objectives such as supporting makes their targeting unfeasible. One option would be innovation, linking the local private sector to global to replace tax holidays with cost-effective incentives value chains of production, training workers, or focusing such as deductions on the acquisition of fixed assets on higher value-added sectors. As an overall guiding for which marginal investors in capital-intense activities principle, incentives should only be used to encourage would “self-select” as beneficiaries, as discussed the private sector to fund public goods (such as capital below. If policymakers deem that tax holidays are still formation or infrastructure) and should be avoided for necessary, their fiscal impact could perhaps be lessened 42 Cambodia: A Public Expenditure Review by reducing the length29 or introducing partial tax example, if one firm producing a certain good qualifies holidays. The latter could take the form of a 50 percent for a VAT exemption while another firm producing CIT tax exemption rather than a total tax holiday, or a the same product does not, the first firm can sell its gradual decrease in exemption rates over time. commodity at a lower price in the domestic marketplace than the second. VAT exemptions at customs may also Cost-effective incentives aimed at boosting distort the VAT chain beyond the importer and result in investment levels could be more efficient in loopholes and unintended foregone revenue. One of the increasing capital formation in Cambodia. As in strengths of the VAT is its consistency; it ensures that all some upper middle-income Asian economies,30 these tax due to the state is paid and reduces opportunities for incentives could take the form of investment tax credits tax avoidance. Removing existing VAT exemptions for (ITC) or investment tax allowances (ITA), consisting in importers while granting them prompt payment of their the deduction of a fixed percentage of capital spending input tax credits would help improve the productivity of against the tax liability (ITC) or the business income (ITA), this tax. with a limit. Similar to other countries in the region such as Malaysia, the amount of the deduction could range It would also be advisable to reduce potential between 60 and 80 percent of the value of the asset loopholes around imports for final consumers/ (e.g. a machine). Liabilities can be carried forward for a individuals. While certain officials and diplomats enjoy few years and ascribed when taxes are paid. One virtue tax exemptions on the import of goods, detailed analysis of these incentives is that they effectively reduce the of imports by individuals would be useful in helping to cost of capital and have a self-correction mechanism prevent potential tax evasion activities (for example, the since low-margin investors (those making little profit in import of vehicles by individuals to sell them to local the first few years) will be more incentivized than large- retailers or final consumers). Tax expenditure on imports margin investors (who may only be able to carry forward by individuals amounted to 0.47 percent of GDP in 2015. deductions over one or two years due to large profits). It would thus be an effective way of targeting marginal Although Cambodia is one of the few countries in investors who are considering coming to Cambodia the region that does not have incentives related to but would need to see the cost of inputs lowered over R&D, such incentives are expected to be effective the first few years as they get established. Another only in the medium term so are less of an immediate advantage of these kinds of deductions compared to priority. Evidence indicates that the rates of return for plain exemptions on corporate income is that firms R&D in poorer countries are usually low due to the would need to fill out tax returns, which would likely absence of complementary factors such as education, result in better taxpayer information and reduced scope scientific infrastructure, functioning of the national for tax evasion. More generally, cost-effective incentives innovation system, and quality of the private sector are expected to help increase the competitiveness and (Goñi & Maloney, 2017). Thus, at least in the short run, level of diversification of the economy by encouraging investing in human capital and improving the overall investments into capital-intense sectors. framework and enabling environment for innovation would likely be more effective than granting incentives To reduce tax evasion, VAT exemptions could be to companies for conducting R&D. In the medium term, replaced with tax credits. The current VAT exemption deductions on R&D and training could be introduced for policy creates distortions in the marketplace. For all sectors (including services). 29  Cambodia (up to 9 years), is on the high side in the region, together with Lao PDR (10 years) and Thailand (11 years), none of which are direct competitors with Cambodia in terms of FDI. Meanwhile, other countries have much shorter holiday periods, such as Philippines (6 years), Myanmar (5 years), Vietnam (4 years) and China (2 years). 30  China, Thailand, Malaysia, and Singapore. Cambodia: A Public Expenditure Review 43 Enhancing the transparency and yearly tax expenditure report or, ideally, as an annex to management of investment incentives the national budget. The legal framework for tax incentives should be Cambodia could also consider streamlining the consolidated and placed in the Tax Code rather approval process for incentives, implementing than in the Investment Law or sectoral legislation. clearly pre-defined and objective eligibility criteria Following international best practices, moving the to reduce room for discretion and uneven treatment tax incentives to the Tax Code would allow taxpayers of investors. Notable efforts have been made to reduce to access all the needed information and prevent delays in the screening and approval of QIP incentives, the proliferation of incentives that could result in tax including the introduction of the 31 working days limit expenditure. Tax authorities are also better positioned by the amendment of the LOI. However, investors are than other institutions to monitor companies’ compliance still required to get certificates of eligibility from different with the incentive regime, given the complementarities institutions. Redundant application processes that of these checks with the primary functions of the tax are essentially duplicated across different reviewing administration. agencies could be eliminated. Based on international experience, the process could be streamlined further This would also be a good opportunity to consolidate with the introduction of automatic approval with ex-post and streamline the information on incentives that is revision by the tax administration, moving to a self- available to investors. The specific criteria used to declaration system instead of the current procedures grant tax incentives and a detailed explanation of the that require authorities’ approval of each individual QIP. approval and appeal process should be made public (and available in both Khmer and English). This could Another potential reform is to move from a project- be undertaken as part of a reorganization of the relevant by-project basis to granting incentives to firms. In websites in a more user-friendly manner, following some cases, the granting of exemptions to QIPs rather international best practices. than firms results in companies continuing the same activity under a different QIP to continue receiving It would be desirable to produce annual tax exemptions beyond the established period. Furthermore, expenditure estimates as an annex to the national in the case of firms carrying out both QIP and non-QIP budget. Tax expenditures are a use of public funds, activities at the same time, no clear regulations exist on and therefore information on them should be made how to separate profits between the two. This obviously available to the public. The tax expenditure estimates creates an incentive for firms to classify all their profits based on customs data presented earlier in this report under the QIP regime. could be updated annually and complemented with further analysis using customs data to trace imported To reduce confusion, the current formula for goods by exporting companies, which could help preferential taxation could be substituted with a confirm some of the findings of this report. This would standard length of time for the incentive (from the also need to be complemented with GDT information date of registration). Interviews with stakeholders on CIT and VAT exemptions. Data could be produced indicate that there is confusion on what constitutes the on the different types of incentives awarded, estimated trigger period for receiving investment incentives and tax expenditure amounts, and beneficiary sectors. The how to interpret it. Replacing the current 3+3+3 years consolidated assessment could be published as a scheme31 with a more straightforward classification of the applicable duration of incentives would provide greater 31  Trigger period (starting from receiving QIPs status, and lasting until the first profit is made or 3rd year after first years of sales) + 3 year-holiday periods + priority period (1 to 3 years depending on the capital invested). 44 Cambodia: A Public Expenditure Review investor certainty and reduce administrative complexity. targeted policy objectives could then be assessed, CIT tax holidays and exemptions in customs duties for including an evaluation of the short- and long-term inputs could be time-bound (for example, to just six benefits and costs of the incentives provided. Incentive years), while other incentives in the form of deductions programs should also include indicators to measures could be open-ended. Over the medium term, tax their results, as well as monitoring mechanisms to holidays could also feature a progressive reduction in make sure public money is being spent effectively. exemption amounts (e.g., declining from 100 percent to For example, if the policy objective is to attract higher 20 percent over six years and to 0 in the seventh year). value-added and capital-intense industries, regular Revenue gains stemming from rationalized tax holidays measurement of capital stock and productivity at the will enable the government to investment in necessary firm level could help assess the effectiveness of some of infrastructure. the incentives (e.g., the ITA/ITCs). This information could then be used to make adjustments if necessary. Cambodia could also consider eliminating the minimum investment threshold required for QIPs Regional and inter-institutional coordination is also and instead introduce tax allowances or tax credit essential to achieving the objectives of investment on capital investment. At the moment, QIPs are attraction and retention. While Cambodia has granted to companies presenting a minimum capital numerous inter-ministerial working groups and task investment. 32 This minimum capital requirement could forces working on enhancing the business environment, be removed for several reasons. First, firms could tend not all of them have proven to be effective. To implement to overstate the amount to be invested in the QIP; in the proposed reform programs successfully, high-level fact, approved QIP fixed assets in 2013 (USD 4.8 billion, (and working-level) officials from all concerned ministries according to CDC) were substantially higher than official and agencies must work together toward a set of FDI net inflows (USD 1.3 billion) for the same year. common objectives. The IDP offers one possible avenue Diligently checking whether the amounts declared as for the government to align around a strategic approach QIP fixed assets are effectively invested would be costly to economic diversification and upgrading, but effective and would require substantial resources and capacity implementation remains a challenge. In addition, in the at the CDC. In addition, this threshold results in firms context of a region where there is a pervasive race to with smaller investments potentially competing at a the bottom in terms of investment incentives, ASEAN disadvantage with respect to larger firms. Instead of could be the platform to negotiate and agree upon a this minimum threshold, a series of tax allowances and set of basic principles in the design of tax holidays in tax credits could be granted to any firm that invests order to avoid unfair competition, limit tax expenditure, in the acquisition of equipment and the upgrading of and maximize investment attraction. The East African facilities, and deductions would be audited by the tax Community and the European Union are examples of administration. tax incentive policy coordination at the supra-national level. It is important for Cambodia to engage in a regular review of its incentives regime. The diagnostic analysis presented in this report is an example of such review. Policy objectives should also be clearly defined and agreed upon across public and private actors, including the type of FDI Cambodia would like to attract. The cost-effectiveness of incentives in achieving the 32  The minimum capital investment ranges from USD 100,000 for suppliers to exporting industries to USD 500,000 in garments and USD 1 million for chemicals, fertilizers, or medicines. Cambodia: A Public Expenditure Review 45 Table 12. Policy options towards an investment incentives reform Tax instrument Short-term (1-2 years) Medium-term (3+ years) CIT tax holidays Limitation of 6 years & simplification Introduction of declining exemption rates Exemption in customs duties for No change List of eligible inputs instead of inputs Qualified Investment Projects VAT zero rate on import of inputs Preparation / piloting Implementation for production Tax allowance/ tax credit on capital Implementation investment Deductions to R&D, staff training No change Implementation Source: World Bank staff elaboration. 46 Cambodia: A Public Expenditure Review 3 STRENGTHENING PUBLIC INVESTMENT MANAGEMENT TO SCALE UP PUBLIC INFRASTRUCTURE Introduction and motivation components in DP financing as discussed earlier, it I declined to 4.7 percent of GDP in 2016. Except for 2010, the level of domestically financed capital expenditure n Cambodia, the provision of public has been more modest at around 2-2.5 percent of GDP infrastructure has been financed mainly by in 2009-16 (Figure 33, right panel). DPs. If development partner (DP)-financed projects are considered, public investment has been Cambodia faces three major challenges in sustaining relatively high compared to neighboring countries an adequate level of public investment. First, (Figure 33, left panel). Externally financed public capital externally or DP financed capital is expected to keep expenditure averaged around 4.5 percent of GDP in declining as a percentage of GDP, or at least the portion 2000-08 and increased in the aftermath of the global financed by grants. Second, the recent expansion of financial crisis, reaching a peak of 7.3 percent of GDP the wage bill as discussed earlier is limiting the extent in 2010. Since then, in the context of decreasing grant to which authorities could increase the domestically Figure 33. Public investment has been relatively high thanks largely to DP financing, although externally financed capital expenditure is on the decline Public investment, percent of GDP Capital Expenditure (% of GDP) 12 12 10 10 8 8 Percent of GDP 7,1 6 6 4 4 2 2 0 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Cambodia Lao PDR Indonesia Malaysia Philippines Thailand Singapore Vietnam Locally financed Externally financed Source: TOFE, Ministry of Economy and Finance, IMF Article IV staff reports, Asian Development Bank, and CIA. Note: Data are for 2015 or latest available years Cambodia: A Public Expenditure Review 47 Figure 34. Cambodia still lags behind its peers in the quality of transportation infrastructure Transportation Infrastructure Quality index (1-7) Paved roads as percentage of total roads 7 82.8 80.4 75.9 6 5 56.9 57 4 47.6 3 28.2 2 10.5 9.9 1 6.3 0 Cambodia Philippines Vietnam Indonesia Malaysia Guatemala Nicaragua Bangladesh Cambodia Philippines Thailand Vietnam (2004-15) (2003-14) (2007-13) (2009-11) (2006-10) Transport 2017 Transport 2010 Beginning period End period Source: World Economic Forum, Global Competitiveness Indicators; The World Factbook, CIA. financed capital budget, although revenue collection infrastructure has not been maintained sufficiently due keeps improving. Third, the relatively high level of to constraints in operations and maintenance budgets public investment with limited infrastructure outcomes to be funded by domestic resources. Fragmentation indicates that enhanced quality of public investment of public resources and parallel budget formulation, management (PIM)—doing more and better with fewer budget execution, monitoring and evaluation, and resources—has not yet been achieved. financial reporting undertaken separately for DP- and domestically funded spending may have contributed to Notably, large DP-driven public investment does not this problem. seem to have resulted in substantial infrastructure improvements in recent years, which points to For Cambodia, improved public expenditure spending inefficiencies. Although total public capital management and long-term growth prospects formation has been substantially higher (an average of will hinge upon enhancing the quality of public 7.5 percent of GDP in 2000-15) than in most ASEAN investment management. In 2011, an index of public countries, Cambodia still lags other structural peers in investment efficiency classified Cambodia among the terms of quality of transportation infrastructure, which bottom 40 percent of developing economies, mainly according to the World Economic Forum, has not due to weaknesses in investment project selection improved in recent years (Figure 34, left panel). While and appraisal (Dabla-Norris, Brumby, Kyobe, Mills, & Cambodia’s physical infrastructure has expanded (e.g., Papageorgiou, 2011). In many low-income countries, paved roads as a percentage of total roads) it also inadequate PIM has been found to hamper the remains far behind regional peers (Figure 34, right panel). contribution of capital spending to growth, as well as Given that DP-funded capital investment accounts for the return on investment (Esfahani & and Ramirez, 2003; three-quarters of total public investment, this may reflect Haque & Kneller, 2008). To overcome this challenge, inefficiency in DP-funded capital investment spending Cambodia would need to not only sustain high public even despite relatively well-managed DP-funded PIM. investment levels but also improve PIM to enhance One of the root causes seems to be that DP-financed value for money. 48 Cambodia: A Public Expenditure Review This chapter provides a systematic assessment linkage between the overall policy objectives under the of the PIM system in Cambodia to identify areas National Strategic Development Plan (NSDP) and the for improvement and specific recommendations Rectangular Strategy, and sectorial policies via annual for reform. The diagnostics are based on the World budgets. However, by 2015 most ministries had not Bank Diagnostic Framework, developed to assess the yet reflected public capital expenditure in the BSP, and efficiency and efficacy of PIM in a country (Rajaram BSPs do not play an important role in decisions about A., Le, Kaiser, Kim, & Frank, 2014). While the chapter the annual or capital budget. Moreover, the Budget focuses on government-financed public investments, Law does not include any specific legislation, chapter, some aspects of Public-Private Partnership (PPP) or article for regulating capital expenditure and PIM. and DP-funded projects are also discussed. Ideally, Finally, while the Procurement Law (enacted in 2012) the PPP framework and reforms would be aligned is functional for PIM—in particular, for the downstream with PIM reform to help ensure a unified approach for project/budget implementation process—it could be integrating these elements into the broader PFM reform further elaborated on to link procurement with other agenda. 33 The remainder of this chapter is structured PIM processes (e.g. planning, ex-post evaluation) more as follows. Section 3.2 reviews the legal and institutional broadly. arrangements for PIM in Cambodia, then the findings of the PIM assessment are presented in Section 3.3. The A Public Investment Program (PIP) Sub-decree36 was chapter concludes with a discussion of policy options passed in 2000 to provide guidance on the selection in Section 3.4. and monitoring and evaluation of public investment programs, but it has shortcomings. According to this Sub-decree, the Ministry of Planning (MOP) and the Ministry of Economy and Finance (MEF), with inputs The Legal and Institutional from the Council for Development of Cambodia (CDC), Arrangements for PIM in Cambodia would issue an Annual Joint Inter-Ministerial circular to guide the preparation of public investment programs Although Cambodia has embarked on a Public based on the results from previous years. In this sense, a Financial Management Reform Program (PFMRP) three-year rolling PIP has been developed and updated since 2004, the PIM function does not have an annually with the aim of coordinating and channeling appropriately defined legal framework. The Public domestically and externally financed budget resources Finance System Law (2008) describes what the capital to priority areas of development. While the PIP Sub- expenditure of budget is to be used for but does decree specifies the institutional arrangements for the not specify the processes for streamlining decisions establishment of PIP, it does not refer specific guidelines of capital budget preparation and execution 34 The or manuals that set out the specific ‘rules of the games’ introduction of three-year rolling bottom-up budget on how project proposals are to be pre-screened, strategic plans (BSPs) and an initial phase (partial) of top- appraised, and selected for financing. Moreover, the down Medium-term Budget Framework (MTBF) 35 have Sub-decree does not provide detail on the various helped improve budget credibility and the budget-policy follow-up stages of the PIM process (project appraisal, 33  It is recommended that there be no distinction for PPP projects from the regular PIM, given that PPP investments are not intrinsically different but only provide a different modality for implementing a public investment project. If PPP investments are planned, appraised, selected, and budgeted separately from traditional government investments, this may create undue fiscal risks and concerns on the appropriate forms of accounting, report- ing, budgeting, and implementation. For further discussion, see Jay-Hyung Kim, Martin Darcy, and Nataliya Biletska (2015), Public Investment Man- agement for Public Private Partnership: Analytical Framework and Assessment Tool, mimeo, World Bank. Refer to Recommendation 5 in section V. 34  Recently, the MEF has planned for a comprehensive review and revision of the law to reflect changes and reforms realized under the PFMRP. An important area of the new law will include specific policies and principles regulating the preparation and management of pubic investments at all stages. For more information, see RGC (2015c). 35  Initial (partial) MTBF captures only domestically financed recurrent spending whereby budget “envelopes” for line ministries are tentatively set. 36  Sub-Decree on Public Investment Program Management. No: 36 ANKr.BK. Royal Government of Cambodia, June 14, 2000. Cambodia: A Public Expenditure Review 49 implementation, adjustment, etc.), and the mandate for investment projects. The Letters clarify the rules for ex-post evaluation by line ministries and reporting back project prioritization with a request that each project must on the status and results of project implementation has include a feasibility study and the rules for enhancing the not been enforced. effectiveness and efficiency of projects, and they instruct the ministries to develop guidelines for the design, The Standard Operating Procedures (SOP) for construction, and maintenance of infrastructure. One all externally financed projects/programs are of the shortcomings of these IMC meetings is that they considered adequate regulatory documents for PIM have not yet been formally legislated and are based on but have room for improvement. The purpose of the bilateral (rather than multilateral) discussions between SOP is to improve the efficiency and effectiveness of the MEF and counterpart line ministries. the RGC’s (under the purview of MEF) management and administration of its portfolio of externally funded Thus, while good progress has been achieved projects by clarifying institutional roles, responsibilities, under PFMRP, the legal and institutional framework and accountabilities, as well as streamlining related appears inadequate for effective management of procedures and interactions between and among public investment—especially for the domestically the institutions involved. The SOP aims to clarify and financed component. An initial legal and institutional streamline those RGC and DP procedures and eliminate review points to the lack of a comprehensive and overlaps so operations are implemented efficiently, systematic PIM framework and the limited integration of according to agreed quality standards, and within externally and domestically financed projects (with parallel budgeted allocations. While DPs assist the RGC with budget management systems). The Public Expenditure financial assistance and expertise as necessary, the and Fiscal Accountability (PEFA) Assessment in 2015 RGC through executing agencies and implementing also noted that the PFM systems do not support policy- agencies leads project formulation, preparation, and budget linkages in resource allocation because strategic implementation processes. SOP documents are sector planning remains limited, which may result in acknowledged to be effective for project identification, suboptimal selection of capital investment projects project preparation, and project implementation in and insufficient consideration of the recurrent budget externally funded projects but include little detail about implications of completed projects (RGC, 2015). The the initial stages of project concept identification and incipient MTBF is not fully developed and integrated into development planning. The Procurement Manual formal budgeting processes. Moreover, lack of detailed complements SOPs and supports the improvement of reporting on the implementation of externally funded effectiveness and efficiency in procurement of externally projects and limited integration of those investments in financed projects. the budget hinder authorities’ ability to track the extent to which resources reach service delivery units. The MEF issued a series of Ministerial Letters during 2014-16 to improve the efficiency of public Global trends suggest that PIM requires an investment (Chapters 21 and 61) in three key line appropriate legal and regulatory framework to ministries that receive most of the domestically ensure it functions well and that common standards financed capital budget.37 For the three targeted line and methods are applied consistently.38 The legal ministries, inter-ministerial committee (IMC) meetings tradition and administrative culture in any given country are held to review and approve new and ongoing capital will determine the most suitable approach, but where 37  Those ministries are the Ministry of Public Works and Transport, Ministry of Rural Development, Ministry of Water Resources and Meteorology. Chapter 21 is for direct investment projects, and Chapter 61 is for rural road repairs and maintenance. 38  See recent World Bank PIM technical assistance reports for Cyprus (2016), Georgia (2016), Romania (2015), and Zambia (2013). Other coun- try-specific assessments have been analyzed and summarized in chapter 3 of Rajaram et al. (2014). 50 Cambodia: A Public Expenditure Review Table 13. Cambodia’s legal and regulatory hierarchies for PIM still have a number of gaps Hierarchy/Projects Government-financed projects Externally financed projects Tier 1: Legal authority Public Finance System Law is not Public Finance System Law is not specific in PIM provisions specific in PIM provisions Procurement Law is adequate de Procurement Law is adequate de jure jure Tier 2: Procedural guidelines PIP Sub-Decree (only planning and SOP, Procurement Manual evaluation) Ministerial Letters for IMC proce- dures adopted recently Tier 3: Methodological manual None No standard RGC version (DPs use their own) Source: World Bank staff elaboration. possible, the guiding principle should be to avoid The Public Finance System Law should be at the top using an overarching framework and legal instruments tier as the primary legislation for establishing the legal to provide detailed procedural arrangements and authority for PIM (Table 13). Unfortunately, the current instructions. Procedures and methods usually have to PIM articles do not explicitly regulate the separate PIM be adapted and changed over time (particularly in the stages of project preparation, appraisal, selection and early days of PIM reform), and the more authoritative a capital budgeting, and project implementation. For the legal instrument is, the harder it will be to initiate such second tier, even though externally financed projects change (IMF, 2007). This principle points to the design are regulated by the SOP and Procurement Manual, and adoption of a hierarchical legal and regulatory PIM domestically financed projects are rarely regulated with framework, consisting of three tiers from top to bottom: formal guidelines. At the third tier, no standard manual for the RGC exists, although some DP manuals are 1. Legal authority for the PIM system which usually is available. established through some form of primary legislation. 2. More detailed procedural guidelines, high-level decision criteria, and designation of analytical tools.39 Diagnostic Assessment of PIM in 3. Methodological guidance, detailed criteria, Cambodia standardized parameter values, and procedural documentation. Generally issued directly by the PIM This assessment is based on a PIM diagnostic tool coordinating agency and in the form of manuals, and has been prepared in close collaboration with templates, and circulars, under the authority granted Cambodian authorities and DPs. The diagnostic to the agency by the primary legislation. framework provides a comprehensive view of the public investment cycle and identifies institutional and Cambodia still has a long way to go in developing procedural gaps across the eight stages of PIM against such an appropriate legal and regulatory framework. functional (so-called “must-have”) features (Rajaram 39  Usually referred to as ‘secondary’ legislation. Cambodia: A Public Expenditure Review 51 Box 3 The World Bank’s PIM Diagnostic Framework The World Bank’s Public Investment Management (PIM) diagnostic tool is broadly consistent with the approach taken in the Public Expenditure and Financial Accountability (PEFA) initiative. The core dimensions of the diagnosis are set out in the desirable features of eight “must-have” stages of an efficient PIM system, such as: 1) Investment Guidance, Project Development, and Preliminary Screening: the extent to which there is sound guidance on national and sector policy priorities, a formal process for project development, and first-level screening of all projects for strategic alignment. 2) Formal Project Appraisal: the quality of appraisals presented to justify new investment spending. 3) Independent Review of Appraisal: the quality of review of project proposals and appraisals. 4) Project Selection and Budgeting: the process of deciding on priority projects to be funded in the budget. 5) Project Implementation: actual construction of the physical assets and budget execution. 6) Project Adjustment: monitoring of project implementation and adjustments as necessary. 7) Facility Operation: use of the assets for service delivery. 8) Basic Completion Review and Evaluation: ex-post collection of data on total cost and time compared to plan, and selective evaluation of project results. Consistency Authority to Maintain asset Evaluation to in project screen and register, operate improve preparation reject projects and maintain asset guidance 1 2 3 4 5 6 7 8 Implementation Independent Adjustment Educación Operation Guidance Appraisal Selection review Link to a Key to An effective budget and development credible procurement process to support strategy selection implementation and operation Source: Rajaram et al. (2014). A. , Le, Kaiser, Kim, & Frank, 2014). The must-have and cooperation of the MEF, involved the collection of features are designed with the intent not to establish diagnostic evidence. In addition to desk review work, a gold standard but to provide a logical and internally surveys based on written questionnaires,40 other consistent system that all countries should try to follow modalities of data collection, and face-to-face interviews to establish basic discipline in managing their capital with officials at different core and line ministries as well budgets (Box 3). This study, undertaken with the support as other DP representatives were conducted. 40  The preparation of questionnaires was guided by the World Bank PIM diagnostic tool (Rajaram A. , Le, Biletska, & Brumby, 2010; Rajaram A. , Le, Kaiser, Kim, & Frank, 2014), with modifications to reflect the country-specific institutional arrangement and decision structure. 52 Cambodia: A Public Expenditure Review Figure 35. Cambodia has some gaps in its PIM institutional arrangements Stage 1: Stage 2: Stage 3: Stage 4: PIM Project Selection Stages Strategic Guidance Project Appraisal Independent and planning Review and budgeting Responsible GDB of MEF MOP, LMs, MEF LMs N/A with LMs, Entity Decision by CM Stage 5: Stage 6: Stage 7: Stage 8: PIM Project Stages Project Facility and Ex-post Implementation Ajustement Maintenance Evaluation Responsible MDAs, GDICDM LMs, DI and Entity MDAs N/A and GDPP of MEF GDICDM of MEF Note: Ministry of Planning (MOP), Line Ministries (LMs), Ministry of Economy and Finance (MEF), None Available (N/A), General Department of Budget (GDB), National Council (NC), Department of Investment (DI), General Department of International Cooperation and Debt Management (GDICDM), General Department of Public Procurement (GDPP), Ministries, Departments, and Agencies (MDAs). Source: World Bank staff, based in discussions with Cambodian authorities. A review of the roles and responsibilities of the Cambodia, the Rectangular Strategy and NSDP, along major players reveals some gaps in PIM practice with a sector plan, are meant to guide and anchor for Cambodia, as shown in Figure 35. Among the the LMs and its agencies to prepare, pre-screen, eight must-have stages of good PIM practice, and appraise the projects. In addition, efforts have stage 3 (independent review) and stage 8 (ex-post been made to institute a cross-framework in policy evaluation) are missing in Cambodia. In the upstream and budget dialogues between LMs and MEF as a decision process (stages 1-4) as well as the downstream collaborative mechanism to support project appraisal implementation process (stages 5-8), the gatekeeping and selection. The above-mentioned strategic oversight role/responsibility of MEF is limited and could planning documents should provide the overarching, be strengthened further. Across the various stages, the first level of policy direction, with the BSPs and PIPs PIM system for domestically financed projects is much flowing into implementation plans. However, both the weaker than the one for the externally financed projects. Rectangular Strategy and NSDP strategies – being These gaps are discussed in greater detail below. strategic documents – are inherently broad, as they contain a large number of objectives with less clearly Stage 1: Investment guidance, project defined outputs and outcomes. In this context, it development, and preliminary screening appears that line agency, local authority, and cabinet priorities compete for funding, making the link between Although strategic guidance for public investment NSDP priorities and Rectangular Strategy decision decisions (both domestically and externally making even less clear. Lack of regulations on how financed projects) is provided to implementing project approval, financing, and execution are to be agencies, the legal framework for Cambodia’s managed has resulted in a major disconnect between capital investment program results remains the national strategies and the annual capital budget disconnected and thereby disaggregates the process that determines actual resource allocation. critical stages in policy, planning, and budgeting. Table 14 shows the key Stage 1 diagnostic questions The quality of strategic guidance and planning for used in the survey and interview processes as well as public investment is critical for success in PIM. In the assessments. Cambodia: A Public Expenditure Review 53 Table 14. The investment guidance and strategic planning stage in Cambodia PIM Diagnostic questions Assessment: Assessment: Domestically financed projects Externally financed projects Is there well-publicized strategic Yes, basic guidance in NSDP, RS, Yes, basic guidance in NSDP, RS, guidance for public investment PIP, BSP, and MTBF. However, PIP, BSP, and MTBF. However, the decisions at the central, ministerial, the strategies remain discon- strategies are broad, with a large and departmental levels? nected, with a large number of number of objectives and without objectives and without definition of definition of clear outputs and out- clear outputs and outcomes. The comes. The connection between connection between the guidance the guidance and specific invest- and specific investment decisions ment decisions is weak. is weak. Is there an established formal Less formal process under weak Yes. process for screening project pro- legal and regulatory requirements. posals for basic consistency with government policy and strategic guidance? Is this process enforced and effec- Weak. Development strategies Yes. tive? have limited authority due to quality issues and frequent changes. Source: World Bank staff, based in discussions with Cambodian authorities. Stage 2: Formal project appraisal Nonetheless, Cambodia lacks a manual or set of guidelines on analytical methodologies for assessing Project appraisal is not applied comprehensively in the feasibility and sustainability of domestically financed Cambodia, with externally financed projects being projects. Rejections of feasibility appraisals are few, subject to systematic appraisal according to DP including for externally financed projects. The diagnostic requirements with domestically financed projects questions related to project appraisal that were used in largely not going through formalized appraisal. For the surveys and interviews as well as the assessments project appraisal to be effective, technical guidance are shown in Table 15. and capacity, funding, and a relatively clear process and allocation of responsibilities should be in place. In Stage 3: Independent review of appraisal the absence of a formal and regulatory allocation of roles within RGC agencies, appraisal for domestically While independent reviews are carried out for financed projects remains largely focused at the level externally financed projects, Cambodia has no of the budget entity promoting the project, where formal arrangements for external review of the internal procedures—largely informal—apply. The appraisal study and decision to proceed for MEF has taken some initial steps to inject the rigor in domestically financed projects. It is good practice to project appraisal by establishing and chairing bilateral subject project appraisal decisions to an independent cross-ministerial steering committees at the policy and review to challenge underlying the assumptions, technical level, and by setting some selection criteria. 41 forecasts, and conclusions of a sponsoring entity. 41  The MEF Department of Investment, for example, highlighted the elaboration of a list of annual preparation priority projects at MPWT. This ministry‘s infrastructure projects are being constructed following formal cross-agency and inter-ministrial review, and they must be approved by the MEF prior to receiving financing. 54 Cambodia: A Public Expenditure Review Table 15. The project appraisal stage in Cambodia PIM Diagnostic questions Assessment: Assessment: Domestically financed Externally financed projects projects Is there a formal pre-feasibility study process before the No formal pre-feasibility No formal pre-feasibility formal appraisal process of feasibility study? process. process, but done so selectively. Is there a formal cost-benefit appraisal process for more No. Yes, at the line ministry detailed evaluation (whether at the line ministry or the level. central finance agency level) of public investment project proposals? If yes, is appraisal mandatory for all projects or only for pro- N/A For all projects, but jects above a certain monetary value? Is project appraisal seemingly selectively undertaken only for specific sectors, and if so, which done across sectors. sectors? What proportion of public investment projects is formally appraised for costs and benefits? Is there a formal manual or guidelines for cost-benefit No. Yes, but dependent appraisal? What are they? What is the quality of such ap- on each DP’s manual/ praisal manual or guidelines? guidelines (e.g., WB/ ADB). There is no RGC version. Are project appraisals formally undertaken by the spon- N/A Both by the sponsoring soring department or by an external agency? What is the department and/or by quality of such appraisals? an external agency. Quality is uneven. What proportion of such appraisals is rejected or sent back N/A Not known (but seem- for amendment? ingly not many rejected cases). Source: World Bank staff, based in discussions with Cambodian authorities. However, this would be difficult in present circumstances Stage 4: Project selection and budgeting due to the currently unstructured way that projects are appraised and financed in Cambodia. Even in externally The assessment found that Cambodia lacks sufficient financed projects, some external reviews of capital cost formal guidelines for selecting domestically financed estimates are done at the detailed design stage and projects as well as formal processes for integrating not at the earlier feasibility stage. Although the MEF the investment and recurrent implications of has introduced a new review and approval process projects. As one of the most contentious stages of through the IMC for three selected line ministries as the investment cycle, with multiple actors potentially discussed earlier, in practice, the rules for prioritization trying to exert influence for political or other reasons, and efficiency improvement contained in the Ministerial project selection is perhaps the most critical stage of Letters remain too general.42 Table 16 summarizes the the investment management process. Project selection appraisal review function in Cambodia PIM. criteria have not been formally established in Cambodia, 42  See a series of Ministerial Letters to the Prime Minister copied to three ministries (Ministry of Public Works and Transport, Ministry of Rural Development, Ministry of Water Resources and Meteorology) during 2014-16 (RGC, 2014-2016). Cambodia: A Public Expenditure Review 55 Table 16. The independent appraisal review stage in Cambodia PIM Diagnostic questions Assessment: Assessment: Domestically financed projects Externally financed projects Is there an established independent No. An IMC review process for Yes. Reviews done by development review process for quality assur- three selected line ministries has partners. ance? been introduced since the 2015 budget, but clear guidelines have not been yet formalized. Is this process enforced and effec- No. Yes. tive? Source: World Bank staff, based in discussions with Cambodian authorities. aside from the establishment of the IMC review process implementation and administration manuals and with prioritization criteria for three key line ministries, as handbooks for externally financed projects, it does not mentioned above. It is also essential that the process of have standardized procedures for domestically financed appraising and selecting public investment projects is investments. Due to this void, LMs independently linked in an appropriate way to the capital budget cycle, develop their own internal procedures to suit their needs even if they run on different timetables. In Cambodia, and preferences, so the guidelines for implementation the budget is approved first, then the LMs prepare the are quite divergent across sectors. project. The capital budgeting process is less effective given that some projects that are political priorities While the recent PFMRP appears to have injected may be included top-down. Moreover, although the more rigor into procurement and financial controls BSP provides a framework for integrating the planning since 2004, some challenges remain. The PEFA 2015 of capital and recurrent expenditures, planning for gave a ‘D’ rating for the dimension of “use of competitive the recurrent expenditure implications for operation procurement methods.” Although the default method and maintenance of newly created capital assets is of procurement is competitive bidding, backdoor inadequate. In addition, externally financed projects loopholes such as certain non-competitive bidding are managed largely outside government budgeting, methods remain open and allow direct contracting. accounting, and reporting mechanisms, which continues While bidding opportunities have become public, other to hamper the RGC’s ability to provide comprehensive important information such as government procurement budget and financial reporting of all resources. Table 17 plans, contract awards, and data on resolution of shows the key diagnostic questions that were used to procurement is not public yet. Statistics from 2012 clarify the status of project selection and budgeting. show that one-third of the contracts were issued using less competitive methods with no public information Stage 5: Project implementation justifying the use of such methods based on the new Law of Public Procurement. In terms of expenditure Similar to the upstream stages of PIM, externally and commitment control, LMs point to challenges such financed and domestically financed projects are as weak absorption capacity, the limited number of regulated differently during implementation, with qualified contractors in the local market, and significant a lack of standardized project implementation delays in cash payment to contractors due to manually guidelines for domestically financed investments. done and cumbersome administrative processes. Table While the RGC has issued the Standard Operation 18 summarizes the diagnostic results on challenges Procedures (updated version 2012) and various project faced in project implementation in Cambodia. 56 Cambodia: A Public Expenditure Review Table 17. The project selection and budgeting stage in Cambodia PIM Diagnostic questions Assessment: Assessment: Domestically financed projects Externally financed projects Are there transparent guidelines No, although an IMC review pro- Yes, but weak coordination and and criteria for selecting projects cess with prioritization criteria was harmonization with domestically with reference to policy objectives established recently in three key line financed projects. at the ministerial level and MEF? ministries. Is there a well-structured formal No, mostly an informal process. No No. Even in externally financed budgeting process with scope to formal consideration of recurrent projects, consideration of recurrent integrate investment and recurrent expenditure implications. implications is weak. implications of projects? Is the budgeting process effective Weak, as some top-down projects Yes. in gatekeeping to ensure that only also are included. appraised and approved projects are selected for budget financing? Is the budgeting ensuring adequate No. Yes, in investment financing, financing for selected projects, but with uncertainty in recurrent including recurrent needs, on com- maintenance and operation budget pletion? funding availability. Are externally financed projects Different rules and fragmented. Different rules and fragmented. subject to the same or different Approximately one-quarter of all Approximately three-quarters of rules for selection and inclusion in projects are financed by national all projects are financed by DP the budget as government-financed budget. funding. Different rules cause projects? unbalanced and less harmonized decision making. Source: World Bank staff, based in discussions with Cambodian authorities. Stage 6: Project adjustment in the cross-ministerial mechanism in project preparation, implementation and monitoring. The Department of In general, project monitoring and auditing are Investment, at MEF, has indicated that these ministries not undertaken in a rigorous or consistent manner are in the process of leading the preparation and filing in Cambodia. Active monitoring is critical to detect of the necessary regulatory documents purported to implementation gaps in a timely manner, determine why enhance project auditing. However, it is worth noting they have occurred, and decide whether adjustments to date any effort to institute internal and/or external to the scope, timeline, and/or costing may be needed. monitoring of project execution remains disconnected In Cambodia, due to the absence of standardized from ultimate decision on onward financing or stopping regulation of project implementation, project monitoring of on-going projects. and auditing practices vary widely across LMs, with only a few showing rigor in project monitoring. Formally, project implementation is subject to both internal and external audit, but the quality and Some initial progress has been made to rectify the credibility of such audits remain low. According to weaknesses in key stages of the PIM. Recently, three the PEFA assessment 2015, only 29 of the 40 LMs have ministries which received most national capital budget – functioning internal audit departments in place. Audit specifically MPWT, MRD, and MTWRM—have engaged reporting remains ad hoc and rudimentary. Cambodia: A Public Expenditure Review 57 Table 18. The project implementation stage in Cambodia PIM Diagnostic questions Assessment: Assessment: Domestically financed projects Externally financed projects Are there any formal guidelines on No centrally developed guidelines, Yes, multiple: Standard Operating project implementation procedures with ad hoc procedures practiced Procedures (SOP), DP Disburse- and management? If yes, are they at LMs. ment Handbook, Project Implemen- followed/enforced? tation/Administration Manual. Is it clear who is responsible within Technical department/Project Man- PMU through Prakas (Ministerial the implementing agency for deliv- agement Unit (PMU). Declaration) based on SOP. ering the project on time, to budget, and to specification? Are implementation plans prepared Often specified in (building) contract Yes. for the project prior to commence- documents, but practice varies ment? across LMs. Are movements in total project Spending limit set through ap- Yes, through data on contract costs tracked against the originally proved procurement contracts, award and disbursement. approved budget? but effective controlling of in-year additional expenditures is lacking. What procurement method is used Competitive as default method, but ICB/NCB/direct contract and for the project? Is the e-procure- exceptions allowed depending on shopping. ment system used? the size and nature of contracts (defined in the Law of Public Pro- curement). E-procurement not yet piloted. Source: World Bank staff, based in discussions with Cambodian authorities. Table 19. The project adjustment stage in Cambodia PIM Diagnostic questions Assessment: Assessment: Domestically financed projects Externally financed projects Do in-year progress reports exist, Formally, yes. Frequency of reports Yes. Frequency: quarterly, semes- and what is their frequency? Have differs across LMs, but generally ter, and annually. they covered both physical and monthly, quarterly, semi-annually, financial progress against plan? and annually. Is there any indication that funding Yes, but there are no standard Yes, through reallocation from allocations for the project were terms and conditions for revision of balance of ongoing project to changed in-year because of pro- the project scope, timing, and cost- emergency rehabilitation project. If gress being better than or worse ing. If adjustments are required, adjustments are required, agree- than expected? agreement needs to be reached ment is to be reached jointly by jointly by the LM management and MEF and DP. MEF. Have there been any in-depth Formally, yes. But the extent and Yes, mid-term review. reviews of the justification for the depth of review vary across differ- project adjustment during imple- ent LMs. mentation? Source: World Bank staff, based in discussions with Cambodian authorities. 58 Cambodia: A Public Expenditure Review Table 20. Facility operation and maintenance stage in Cambodia PIM Diagnostic questions Assessment: Assessment: Domestically financed projects Externally financed projects On completion, were the capital as- Yes, inventory made, especially on Yes. sets created by the project added infrastructure like land, buildings, to an asset registry? and roads. The provision speci- fied in organic law (Law on Public Finance System on asset registry). Has adequate funding been pro- Generally, no, although O&M has No. DPs provide capital budget, vided in the budget for O&M of the been given increased attention but the government is responsible new assets created by the project? since 2016. for O&M budget. Have there been any surveys of the Yes, but practice varies across Yes, part of project performance level of use of the new assets by LMs. evaluation, carried out by DP team. the intended recipients? Source: World Bank staff, based in discussions with Cambodian authorities. In terms of project adjustments, a stark contrast prior to project handover, a project completion report can be seen between the rigorous regulatory is prepared and submitted to the executing agency framework for budgetary management and the (EA)/PMU for approval. Once the engineer confirms practically lax control of expenditure adjustment technical completion, the contractor hands over the during construction. Despite strong central control of project to the EA, and the facilities are transferred to the budget, annual expenditure adjustments are both the operating agency. However, the existing disconnect frequent and sizeable. The LMs, in particular, routinely between capital and current budgeting (dual budgeting) tap into the unallocated resources in Chapter 9 of the as discussed earlier often translates into insufficient budget, which are intended primarily for disaster relief funds for operating facilities. The assessment of facility and/or other emergencies. Detailed budget execution operation and maintenance practices in Cambodia is data are not available or maintained by the MEF to help summarized in Table 20. determine to what extent project design adjustments can be justified due to changes in macro or micro Stage 8: Basic Completion Review and local economics or simply due to poor execution and Evaluation monitoring. Table 19 summarizes the key assessment results on project monitoring and adjustment in As in earlier project stages, ex-post evaluation Cambodia. requirements differ between externally and domestically finance projects. Project evaluations are Stage 7: Facility operation important for ensuring that feedback and learning from projects can be used to improve the quality of PIM in While the process for project handover and asset registry the next cycle. In Cambodia, evaluation is conducted for has been established in Cambodia, the disconnect externally financed projects in accordance with specific between capital and current budgeting processes DPs’ project cycle management procedures. However, results in insufficient funds for operating facilities. After ex-post evaluation does not occur for domestically project construction is completed, an effective process financed public investments. Because Cambodia does for handover and the institutionalization of accountability not have effective ex-post evaluation instruments to for effective facility operation are needed. In Cambodia, ascertain whether financing of public investment is being Cambodia: A Public Expenditure Review 59 used efficiently and effectively, the existing procurement role of MEF in upstream decision making for project pre- framework cannot warrant value for money in capital appraisal, appraisal, project prioritization, and budgeting; projects. upgrade the capacity of MEF and LMs to prepare and assess more effective and efficient investment projects that are directly connected to their strategic priorities in the BSP and MTBF; and establish standardized Prioritized PIM action plan guidelines and effective cost controls in downstream project implementation. The recommendations are As described above, the PIM assessment reveals intended to serve simply as guideposts and would need several weaknesses in the system, especially to be prioritized further in accordance with Cambodia’s in the upstream stages of project preparation, specific institutional setting, incentives, and capacity pre-screening, and appraisal, as well as in constraints. Expected timelines are provided based on the downstream stage of ex-post evaluation. experience from global PIM reforms. As discussed earlier, the attribution of roles and responsibilities among key stakeholders is not defined • Recommendation 1: Upgrade PIM strategy clearly in the Public Finance Law and secondary and planning with a PIM sub-decree (Year 1). regulation. The PIM system for domestically financed Such upgrading activity will be based to the extent projects is much weaker than the one for the externally possible on the existing guidance and principles financed projects. For domestically financed investment of the SOP.43 The sub-decree could help improve projects, the absence of a centrally developed, uniform the forward-looking strategic guidance for public set of guidelines, manuals, and templates for preparation investment and strengthen the connection between of project proposals, appraisal, and monitoring of the national strategy and sector planning in the execution drive the vastly different practices at LMs. NSDP and RS with the budgetary framework of the For government-funded public investment, the absence BSP and MTBF. Specifically, the sub-decree would of an independent review of project documents and clarify the principles and decision criteria for project appraisal functions hampers MEF’s role in managing pre-appraisal and appraisal, prioritization, selection, and controlling the allocation and use of capital budget and budgeting for all public investment projects, resources. Downstream, project implementation and including PPP.44 It would also specify roles and operations have shown some progress, with more responsibilities of the MEF, MOP, sector ministries, rigorous procurement and financial controls introduced and implementing agencies at each stage (i.e. project since 2004 in the context of PFMRP implementation. owner, appraiser, reviewer, implementer, decision Nonetheless, there is still room for further improvement. maker), strengthening the gatekeeping role of MEF Finally, Cambodia lacks effective ex-post evaluation as well as the specialized roles and accountability of instruments to assess the efficiency and effectiveness sector ministries to perform necessary assessment of public investment financing. and implementation of public investment projects. The sub-decree should be prepared in the immediate These diagnostic results point to a multi-faceted term and is expected to take 1-1.5 years. set of 11 broad recommendations for short- and long-term reforms in all stages of PIM. The reform • Recommendation 2: Strengthen project pre- recommendations aim to strengthen the gatekeeping appraisal and appraisal procedures and 43  It would be propagated as part of the Public Finance System Law, probably under Article 23 and/or Article 44. Developing such guidelines under the primary legislation is consistent with other country cases of good PIM practices observed in South Korea, South Africa, Chile, Australia, New Zealand, and the United Kingdom. 44  This would comprise the following: appropriateness and relevance in the context of national and sector strategies; economic, social and envi- ronmental viability (through cost benefit or cost effectiveness analysis); affordability and financial sustainability; and implementation arrangements. 60 Cambodia: A Public Expenditure Review guidelines (Years 2-5). Regulated requirements for can understand the recurrent budget implications project pre-selection and appraisal—for example, of the project. A capital budget baseline of ongoing through a PIM Standard Appraisal Manual—are public investment projects for 3-5 years should needed to establish a rigorous national procedure for be established within total resource envelope/ all public capital investment projects. Development constraints, with the total investment budget of the Standard Implementation Manual could be being subdivided into ministry/sector-level budget based partly on the existing SOP for DP-funded ceilings or estimates for capital expenditures. These capital projects. 45 Procedures should contain an recommendations can be developed as part of the explicit description of roles and responsibilities, with PIM operational guidelines and accompanied by a strong demarcation between project sponsors and criteria for project appraisal, review, and selection. decision makers to foster objective decision making. Following international best practices, different • Recommendation 5: Frame the PPP upstream appraisal methodologies and rules tailored to the under a unified framework with PIM (Year scope, timing, and technical complexity of project 2 onward). The Asian Development Bank has proposals should be developed and disseminated. been supporting the RGC in the elaboration of a 46 This strategic policy area can start early but is Concessions Law and a PPP framework—for this expected to continue over the medium term. reason, as agreed with counterparts, this chapter has not looked in detail at PPPs. Nonetheless, the • Recommendation 3: Strengthen independent forthcoming PPP platform should be developed review and oversight (Year 2 onward). Options for in conjunction with, and on the basis of, the PIM institutional arrangements for independent oversight guidelines being drafted under the World Bank and verification of project feasibility studies and TA program. A unified framework will help avoid appraisal should be explored. A two-tier approach undue fiscal risks from fragmented PPP investment is recommended: (i) strengthening the oversight decisions and will formalize and align the PPP role of the MEF through the existing IMC meetings upstream processes of project preparation, pre- and (ii) institutionalizing an independent review appraisal, appraisal, and prioritization/selection with function, which could be completed in 2-5 years PIM. The guidelines for PPP upstream decisions can depending on technical preparation and buy-in from be incorporated into the PIM guidelines.47 government sectors. • Recommendation 6: Strengthen project • Recommendation 4: Upgrade project implementation and procurement (Years prioritization and budgeting procedures (Year 2-5). Standardized project implementation 2 onward). Specific criteria are needed for the guidelines—for example, through a PIM Standard selection and prioritization of projects (see Box 4 for Implementation Manual—are needed to provide examples). Information on capital budgeting could uniform procedures for managing project execution be improved further to include estimates of operating at all LMs. Stricter limitations on the use of non- costs upon project completion so decision makers competitive procurement procedures are also 45  The fundamental steps and processes are provided in the existing Procurement Manual and SOP Manual currently applied to externally financed projects/programs. These have also been used by some government-funded projects in selected LMs such as MoEYS, so the experience of MoEYS could be considered in the development of the new Standard Implementation Manual. 46  For example, Ireland (2005) developed and applied five differentiated appraisal methodologies, ranging from relatively simple appraisal of small projects to full-fledged CBA of projects of large scale and/or highly complex technology. In Cambodia, special rules could apply to small projects (i.e. below a certain threshold, a simple cost-effectiveness method could be used), and some sectors could be excluded from appraisal requirements (e.g., defense, emergency/disaster, State-Owned Enterprises) as most countries do not appraise the entirety of their investment projects. 47  For further elaboration of the unified approach to PPP as an integral part of overall PIM, see Rajaram et al. (2014) and Kim, Darcy, and Biletska (2015). Cambodia: A Public Expenditure Review 61 Box 4 Examples of Prioritization Criteria for Project Selection in Advanced Systems In considering the allocation of funding to new projects, new projects must be verified as having been prepared well and carefully appraised. In more advanced systems, prioritization of new projects is done by spending ministries within sector expenditure limits approved by the government based on prioritization criteria and technical advice from the finance ministry. Some prioritization criteria include: • Alignment of the project with the Government’s expressed national spending priorities and with sector investment plans. • Priority the Government attaches to improving services for the target beneficiaries of the project. • Consistency of the proposed spending with the balance of spending among different sub- sectors within its sector. • Indications that the project will deliver better value for public money than competing and comparable projects as indicated by feasibility studies. • Readiness for implementation, including land acquisition and completeness of detailed designs, implementation plan, bid documents, and procurement plan. • Assuredness of any co-funding for projects funded from multiple financing sources. • Managerial capacities of the implementing agency to deliver the project according to plan. • Fiscal space for project implementation over the medium term, i.e. consistency of the proposed implementation timetable with commitments for ongoing projects and sector/organization capital expenditure limits. • Adequacy and reliability of provision for operating and maintenance expenditures upon completion of the project. needed, and uniform cost standards/format and of the monitoring process, and put pressure on the effective cost controls in budget execution and spending agency to improve quality and timeliness. project implementation should be established. While Box 5 provides an example from the Philippines. developing standardized project implementation guidelines within one year is feasible, the capacity • Recommendation 8: Develop clear project and buy-in of sectors and agencies need to be adjustment procedures and guidance (Years 2-3). built to ensure acceptance and application of these Stronger project adjustment rules and procedures guidelines. should be developed and monitored across the major stages of PIM implementation. Project • Recommendation 7: Establish effective adjustments during construction should be linked mechanisms for monitoring project directly to project results specified in the project implementation (Years 3-5). Experience in other implementation plan and should be monitored with countries points to emerging and unique approaches the budget/funding plan. to project monitoring through IT and Open Government platforms that broaden stakeholder • Recommendation 9: Introduce completion participation, enhance transparency and objectivity review and ex-post evaluation (Year 3 or at later 62 Cambodia: A Public Expenditure Review Box 5 The Philippines Open Roads Platform: a tool for monitoring implementation progress Increasingly, new technologies have been applied in monitoring the progress and quality of infrastructure projects. In the Philippines, for example, the government introduced an online interactive Open Government multi-media platform to digitally map the entire national and local roads network and track associated priority public investment projects. The Open Roads platform leveraged three major technological and practice developments in the Philippines: (i) geographical mapping and geo-tagging; (ii) mobile, drone, and satellite road imaging (photo + video); and (iii) project tracking information systems. The latter interfaces with government systems including the Philippines Government Electronic Procurement System (PhilGeps) and the Department of Public Works and Highways (DPWH) electronic Project Life Cycle (ePLC) database. Open Roads transforms fragmented data into timely information and sends a clear signal through big geo-image data consolidation. More timely and better information across the public investment delivery chain has helped promote better planning, budgeting, procurement, implementation, and participatory audits. The platform has also catalyzed a coalition of engaged national, local, private sector, and civil society stakeholders and ordinary citizens to demand greater transparency, accountability, and participation. Source: The Philippines Open Roads Platform: Promoting Roads to Prosperity with online information and feedback (World Bank, 2015). stage). An analytical project completion report changes and performance. The CAPEX would be can become a requirement of the PIM system. In managed centrally by the MEF. Rigorous control the longer term, a system for undertaking ex-post over the entry and modification of data is needed, reviews of selected major projects (i.e. large-scale and changes should be traceable to individual projects or technically complex projects at the design authorized system users. and implementation stages) could be established. The introduction of completion review and ex-post • Recommendation 11: Design a capacity building evaluation is something that can be sequenced in program for PIM (Years 1-5 and continuous). the medium term. Continuous capacity building can help raise the quality of project preparation, pre-appraisal, • Recommendation 10: Build a PIM database appraisal, prioritization, and budgeting, reducing (CAPEX) to register all relevant information the time and resources required for PIM and on public investment projects and processes capital budget execution and ensuring that public (Years 2-5). A comprehensive official record of all investment projects are prepared and budgeted to data associated with projects during their entire life a higher standard. A capacity building program for cycle, from entering the preparation pipeline through MEF, MOP, LMs, implementing agencies, and other their completion and registration in the central concerned public-sector officials should be designed asset registry, should be developed. Maintaining and implemented to provide training, technical both baseline and actual data sets for each major assistance, and capacity building at all stages of the category for 2-3 years would enable monitoring of project cycle. As a first step, capacity at MEF should Cambodia: A Public Expenditure Review 63 be strengthened given MEF’s proposed central role capital expenditure decision making with that for budget as gatekeeper to screen and review all projects, strategic plans and the annual recurrent budget and including proposed financing. MEF is de facto implementing standardized PIM development and already serving as gatekeeper for most projects, operational guidelines will help ensure that adequate but this should be formalized in the legal framework, budget is available for capital projects. Furthermore, with functions being clearly defined and associated efforts under the current PFM reform to broaden capacity building being provided. Equally important, program budgeting to capture investment projects and sector ministries and agencies primarily tasked reestablishment of the MTBF will further promote the with project preparation and appraisal will need to integration of current and capital budgeting. enhance their knowledge of modern assessment tools and techniques, procedures for analyzing projects, and assessment and management of risks. Ultimately, these reform actions are essential next steps but cannot generate impact on the ground if they are not practical or if discipline and accountability in their implementation are lacking. Reforms are summarized and prioritized strategically over the next five years in Table 21. These guidelines and tools need to be practical, in the sense that they are politically and institutionally compatible. The MEF will lead the development of such missing guidelines and tools in close consultation with all major stakeholders, especially the key LMs. These guidelines and tools should be tailored to the context of the existing PIM and not necessarily rigidly follow all elements of international best practice as observed in other developed PIM systems. Establishing some basic discipline and accountability will also be critical. The most direct way to do so is to institute a mechanism linking project financing (in tranches or installments) to verifiable project implementation results at each critical juncture of the execution period. Once the new PIM framework and system are in place, reforms would need to be harmonized with those of the PFM reform program to facilitate integration of the current dual budgeting system. Once the new, appropriate institutions and procedures for managing the entire PIM system—from project pre- screening through ex-post review—are in place, all externally and government budget-financed projects should follow the same system and processes. Unifying 64 Cambodia: A Public Expenditure Review Table 21. Policy options and Prioritized Action Plan for Public Investment Management Areas for PIM Reform Strategic Suggestions for Re- Prioritized Action Plan forms (i) Strengthen PIM upstream Recommendation 1: Upgrade PIM Year 1: develop a PIM sub-decree project appraisal/selection strategy and planning with a PIM sub-decree Recommendation 2: Strengthen Year 2-3: prepare a PIM Standard project pre-appraisal and appraisal Appraisal Manual procedures and guidelines Recommendation 3: Strengthen Year 3-4: conduct a pilot in infra- independent review and oversight structure to test the developed appraisal guidelines. Annual review and update of guidelines-manual Recommendation 4: Upgrade project prioritization and budgeting procedures Recommendation 5: Frame the PPP upstream processes under a unified framework with PIM (ii) Strengthen PIM downstream Recommendation 6: Strengthen Year 2-3: Prepare a PIM Stand- project/budget execution project implementation and pro- ard Implementation Manual that curement provides uniform procedures for the management of project execution at all LMs Recommendation 7: Establish ef- fective mechanisms for monitoring project implementation Recommendation 8: Develop clear Year 2-5: Establish uniform cost project adjustment procedures and standards/format and effective cost guidance controls in budget execution and project implementation Recommendation 9: Introduce Year 3 onwards: introduce comple- completion review and ex-post tion review and monitoring evaluation (iii) Upgrade PIM database and Recommendation 10: Build a PIM Year 2-5: Develop template and for- build capacity database (CAPEX) to register all rel- mat of the CAPEX. Establish rules evant information related to public and regulations that designate a investment projects and processes unit (sub-unit) to develop, maintain, and update CAPEX Recommendation 11: Design a Year 1-5: Undertake training needs capacity building program for PIM analysis, planning, and implementa- tion of capacity building Source: World Bank staff elaboration. Cambodia: A Public Expenditure Review 65 66 Cambodia: A Public Expenditure Review PART 2 Sectoral reviews Cambodia: A Public Expenditure Review 67 4 INCREASING STEWARDSHIP IN EDUCATION SPENDING Introduction and motivation respectively, in 2015 remain well below the Southeast C Asian averages of 66, 89 and 57 percent, despite recent gains. This is made worse by high dropout rates: only 41 ambodia has done well in expanding percent of those enrolled in secondary education and primary enrollment and now faces the only 27 percent of those enrolled in upper secondary challenges of continued expansion of pre- education complete the necessary grades. Dropouts primary and secondary education. Over the past two have actually risen, resulting in a decline in the lower decades, Cambodia has succeeded in fostering primary secondary education completion rate from 49 percent education enrollment, increasing from a net enrollment in 2008 to the current levels. Consequently, Cambodia rate of 82 percent in 1997 to 94 percent in 2015, a has the lowest share of adults with lower secondary level comparable to high-income economies (Figure education among ASEAN economies (CDRI, 2013) and 36). However, gross enrollment levels in pre-primary, its peers. lower and upper secondary at 19, 54 and 24 percent, Figure 36. Primary education is now universal, yet challenges in pre-primary and secondary education enrollment persist Enrolment rates by level 120 100 80 60 40 20 0 Cambodia Southeast Asia Lower middle-income High-income Pre primary, % gross Primary, % net Lower secondary, % gross Upper secondary, % gross Source: Cambodia; MoEYS Education Statistics 2015-16, Comparator countries and Cambodia pre-primary data; UNESCO Institute of Statistics, 2015. 68 Cambodia: A Public Expenditure Review Figure 37. Proficiency levels in reading and math are generally very low, and progress has been mixed Grade 6 Khmer proficiency levels Math grade 6 test scores, 2007-13 for reading and writing, 2013 580 Reading 560 540 Equated score 520 Writing 500 480 460 Overall 440 420 0% 50% 100% Small rural Large rural Small urban Large urban Below basic Basic Proficient Advanced 2007 2013 Source: EQAD grade six assessment, 2007, 2013. The gains from achieving universal primary next plan, this chapter assesses key challenges in education are undermined by poor learning education spending. Section 4.2 assesses the current outcomes, and progress toward improving learning allocation of public spending in education in relation to outcomes has been mixed. A national assessment of sector priorities and emerging challenges and examines student achievement found that 39 percent of grade 6 how funds reach schools and how they are used. Section pupils had a below basic proficiency rating for reading 4.3 evaluates whether access to education is equitable in Khmer, and 76 percent were rated below basic on and whether spending is pro-poor, and Section 4.4 writing (Figure 37, left panel). Pupils also performed explores whether further gains can be made in terms of poorly in math, with an average test score of 43.4 expenditure efficiency. Section 4.5 then describes the percent. Compared to 2007, test scores have declined factors that affect the quality of education and learning in small rural areas, stagnated in large rural areas, and outcomes. The chapter concludes by presenting policy mostly improved in urban areas (Figure 37, right panel). options in Section 4.6. These poor learning outcomes are echoed in employer’s concerns about recruits who have poor reading and numeracy skills. Program budgeting and flow of Attaining universal high-quality education will be funding to education crucial to Cambodia’s success as a middle-income economy. Aware of the challenge, authorities have in Increased government spending on recent years developed an Education Management education has been driven by higher and Information System (EMIS) to monitor sector teacher salaries performance and future planning, implemented a National Assessment System test, and significantly In recent years, Cambodian authorities have given increased the allocation of funding to the sector. To greater attention and resources to the education support the implementation of the current Education sector, although public spending in education still Strategic Plan (ESP) 2014-18 as well as to inform the lags behind comparator countries. Expenditure Cambodia: A Public Expenditure Review 69 Box 6 Public Expenditure Review 2011. Summary of education recommendations and current status The Public Expenditure Review from 2011 highlighted several issues in the education sector. This box provides a brief overview of identified challenges, as well as current status of the recommendations proposed in the report. Increased educational spending – in line with fiscal space considerations – was recommended (fully implemented). Education was identified as funded in line with the contemporary ESP, but far below countries with broadly similar GDP per capita, at 2.4 percent of GDP including DPs. As such, increased educational spending – in line with fiscal space considerations – was recommended. The RGC has subsequently, from 2010-2015, increased education spending roughly in line with the ESP 14-18, both as a percentage of GDP, from 1.6 to 2.1, and in percent of government budget, from 12.6 percent to 13.9 percent. This has brought Cambodian Education expenditure closer to comparator countries, but still well below the lower middle-income average. A shift in expenditure towards secondary education was also recommended due to a reduction in primary level students (fully implemented). Increased focus on secondary level was recommended to address rising numbers of secondary students. By 2015, this already seems to have been taken into account, with secondary level receiving 37 and 34 percent of RGC and DP funding respectively, both even exceeding stipulated ESP resource requirements. Specific allocation recommendations provided by the PER 2011 were firstly, the need for increased access to Early Childhood Education (partially implemented), citing the benefits of pre-primary school years. While not dramatic, enrolment in ECE has increased, reported at 41 percent in 2016. However, this progress has been DP driven and quality of ECE remains a major challenge. Secondly, the textbook budget was identified as underfunded and in need of increases (partially implemented). It was recommended to increase the budget by at least 25 percent. The budget has remained relatively stable as a percentage of GDP, increasing only from an allocation of 0.036 to 0.04 percent of GDP between 2011 and 2015 (PB activity line 1.8.3). Thirdly, the need for expanded scholarship arrangements was emphasized (fully implemented). Another area where major progress has been made. From only a few scattered pilots in 2010, the RGC co-developed and took over responsibility for a major scholarship scheme in 2015 and has since increased both the number of recipients and benefit levels. In 2015 the total number of recipients, in the government scheme alone, was 150,644 students across primary and secondary education levels and the RGC expenditure had increased from near-zero to 10.9 million USD, or 0.06 percent of GDP. In the medium term, the PER 2011 emphasized lagging quality improvements and the necessity of strategic planning (ESP), with a strengthened feedback process determining target achievement and identifying performance and non-performance (partially implemented). To allow this, development of a strong quality measurement system through student assessments was recommended, as a key data collection mechanism. National test scores have been providing a solid basis for assessing student performance. The development of EMIS is also showing promising progress, allowing for analysis to inform strategy and allocation adjustments. The quality of education remains low, however. According to the World Economic Forum, quality of education remained stagnant for most of the past decade, prior to increasing from 3.15 in 2015/16 to 3.5 in 2017/18, compared to ASEAN and lower middle-income country averages of 3.84 and 3.44, respectively. Nonetheless, education quality and learning outcomes remain a challenge. 70 Cambodia: A Public Expenditure Review Box 6 Other aspects were noted, while no specific recommendations were presented in the PER 2011. First, the issue of teacher absenteeism was noted as a concern and connected to the low wage levels received by teachers. Absenteeism remains a key issue and while minimum wages have been increased from around 80 USD per month in 2012 to 160 USD in 2015 (with a target of 250 USD/ month by 2018), it remains to be seen whether this alone will reduce absence. Teacher shortages and surpluses were also highlighted, a daunting challenge that is discussed in this report. The wage share in expenditure was noted as high, but not problematic, at ¾ of total education spending. It had further increased to 81-88 percent by 2016. DP dependence for Technical assistance was noted as a cause for concern, noting that it had “declined much less than targeted under the ESP 2006-2010”. This dependency remains, with DPs financing 69 percent of all non-wage expenditure and across most policy relevant areas in 2015, including capital expenditures, quality improvements, sector policy and teacher training. Finally, it was noted that while the Public Financial Management Reform was progressing reasonably well, lack of autonomy for budget entities, especially with regards to SOB transfers, remained a problem causing low execution rates. The disbursement systems and autonomy have subsequently been improved significantly, with moves to bank transfer disbursements and introduction of increased flexibility in SOB allocations. Room for improvement remains however, as disbursement procedures remain lengthy in many cases. in education as a share of total public expenditure even if DP funding (0.6 percent of GDP in recent years) increased from 12.6 percent in 2010 to an estimated 15 and NGO funding (0.4 percent of GDP) are taken into percent in 2017, which is similar to the level observed account (Figure 38). in other lower middle-income economies. Government spending through the Ministry of Education, Youth, The increase in public education expenditures has and Sports (MoEYS) increased from an average of 1.6 been driven mainly by higher public-sector wages. percent of GDP in 2010-12 to close to 2.1 percent of Between 2012 and 2016, government spending in GDP in 2015, and was budgeted at 2.8 percent of GDP education more than doubled in nominal terms, mostly in 2017. The Ministry of Labor and Vocational Training driven by an increase in public sector wages (Figure 39). (MoLVT) spending on technical and vocational education During this period, the number of teachers increased and training (TVET) is estimated to have amounted by 3.6 percent, while teacher salaries increased from to 0.05 percent of GDP in 2015. 48 Yet despite these about USD 80 a month in 2012 to USD 160 in 2015, increases, public education spending in Cambodia in line with the increases in the public sector minimum remains below the 4.2 percent of GDP spent on average wage. The minimum wage has since increased further, by low-income and lower middle-income economies, surpassing USD 200 per month in 2017, with plans to 48  MoLVT TVET allocation was taken from MoLVT Program 2: “Development of technical and vocational training and education.” However, actual allocations to TVET may be higher, as salaries for provincial departments have been lumped in a general line department support program. TVET spending was calculated by assuming that TVET received a share (33.1 percent) of MoLVT-executed expenditure equal to that in the budget. Cambodia: A Public Expenditure Review 71 Figure 38. Public spending in education has increased significantly but remains below that of comparator countries Education expenditure, % of GDP 7 6 5 4 3 2 1 0 Bangladesh Cambodia Cambodia Cambodia Cambodia Lao PDR Low-income Lower Thailand Vietnam 2012 2015 2016 2017b middle-income Government education spending Donor-financed expenditure Source: Cambodia: MEF TOFE, CDC Database. Comparator countries: WDI. Note: b. Cambodia 2017: RGC budget and assumed DP funding. For comparator countries, government education spending includes transfers, latest available year. Figure 39. The increase in public sector Figure 40. Although Cambodia has a wages has driven the rapid increase in relatively low proportion of teachers, Government spending on education it spends a relatively large share of its education budget on salaries MoEYS Expenditure by classification, 2010-2016 (Million constant 2000 KHR) 1,000,000 3.0 90 80 2.5 800,000 Teachers per 100 inhabitants 70 % personnel spending 2.0 60 600,000 50 1.5 40 400,000 1.0 30 200,000 20 0.5 10 0 0.0 0 C. Asia Cambodia Low-Income Africa Asia & Pacific W. Hemisphere Mid. East & Middle-Income High-Income Europe 2010 2011 2012 2013 2014 2015 2016 Goods (ch. 60) External services (ch. 61 & 62) Personnel (ch. 64) Subsidies and social assistance (ch. 65 + 62) Taxes and excises (ch. 63) Education personnel per 100 inhabitants Capital expenditure % personnel spending in education budget Note: Data in constant 2000 KHR, using the WDI GDP deflator. Source: Cambodia: 2016 budget. Comparator countries: IMF, FAD, Source: 2010-2014 capital expenditure from MEF TOFE, recurrent and Evaluating Govt Employment and Compensation. all 2015-16 expenditure data from MEF Treasury Accounts. 72 Cambodia: A Public Expenditure Review surpass USD 250 in 2018 as discussed earlier. As a funding is expected to dwindle, and the government result, Cambodia now spends a relatively large share will need to increase its stewardship of non-wage of the education budget on salaries at 81 percent49 expenditures in the same way they have recently taken compared to 69 percent for low-income countries, while over the provision of scholarships from DPs. it lags behind in the proportion of teachers (Figure 40). When DP resources are added, the overall budget Meanwhile, DPs continue to fund more envelope is nearly sufficient to cover the funding than two-thirds of non-wage expenditure requirements spelled out in the Education Strategic in education Plan (ESP), while falling short in capital expenditure. The ESP is the key medium-term planning instrument DPs and NGOs contribute significantly to the funding which details spending requirements associated with and operation of the education sector. DPs fund 69 sector and sub-sector objectives, thus providing a road percent of all non-wage expenditures in the education map and a benchmark for assessing progress and sector, including cross-cutting functions such as performance. Despite the overall increase in education construction of schools, quality improvement programs, expenditure, the allocated government budget sector policy formulation, and teacher training (Figure remained below ESP requirements in 2014 and 2015, 41, left panel). In addition, 180 NGOs, registered in the although DP and NGO funding helped bridge the gap database of the Council for Development of Cambodia, (Figure 41, right panel). Although Government-funded carry out education-related activities, from running capital expenditure in education has increased in recent primary education and TVET centers to the provision of years, it remains modest (0.036 and 0.039 percent school facilities and materials. As discussed earlier, as of GDP in 2015 and 2016 respectively). In 2015, the Cambodia becomes a middle-income country, external combined government and DP expenditure amounted Figure 41. DP and NGO spending helps meet spending goals in the Education Strategic Plan DP spending, share by category Funding requirements and actual 2,500,000 School and Facilities 22.2% Sector Budget Support 14.9% 2,000,000 Secondary Education Teacher 4.2% 1,500,000 Training 1.6% Other 3.8% Sector policy 1,000,000 8.0% 500,000 Primary/Ba sic 26.6% 0 Tertiary, Vocational 2014 2014 2014 2015 2015 2015 and Higher 18.7% ESP Req Budget Exp ESP Req Budget Exp ESP Resource reqother NGO Expenditure DP Expenditure Mo EYS Expenditure DP & NGO Pledges RGC Budget ESP MoEYS resource req. Note: ESP Req: Resource requirement identified in the ESP, Exp: actual executed expenditure. Source: ESP, MEF TOFE, MoEYS Annual Operation Plans, CDC database; expenditure data from MEF Treasury Accounts. 49  In addition, 36 percent of the non-wage component was spent on social benefits and per diem in 2016 which, if added to salaries, would amount to 87.5 percent of total education budget for that year. Cambodia: A Public Expenditure Review 73 Figure 42. Cambodia’s share of spending on higher education is much lower than in comparator countries Tertiary Education expenditure (% share of Government Education expenditure) 40 35 30 25 20 15 10 5 0 Bangladesh Cambodia Cambodia Cambodia Indonesia IGuatemala Malaysia Nicaragua Philippines Thailand Vietnam pro-rata MEF budg MEF exp Note: Pro rata is the term used to describe a proportionate allocation. It is a method of assigning an amount to a fraction according to its share of the whole. The method was utilized by RGC (2015) to estimate sub-sector expenditures, using a combination of students and staff numbers to determine the share of non-specified expenditure. It is included here as an upper estimate of possible HE spending. Source: Cambodia: MoEYS 2015 Budget and Treasury accounts, SNEC pro-rata calculations (highest estimate) in red. Comparator countries: UNESCO Education Dataset (June 2016 release). to 0.21 percent of GDP and covered just 57 percent of Significant efforts have been made to the capital expenditure requirements envisaged in the improve program budgeting and the flow ESP. There are promising signs: the RGC increased its of funding to schools, but challenges capital budget for 2017 to KHR 90 billion, nearly tripling remain the budgeted KHR 35 billion in 2015. However, while the increase is dramatic in relative terms, the allocations Significant progress has been made in introducing are still low in absolute terms. program budgeting (PB) in education, but more remains to be done to make it a meaningful instrument Overall, the distribution of funding through of policy implementation. PB was first piloted at the different education levels seems well-aligned with MoEYS in 2007 and was formally introduced in 2015 as ESP requirements, except in the case of tertiary part of PFMRP Phase III, with the aim of improving the education. Cambodia’s higher education share in total linkages between budget and policy outcomes. Despite spending is significantly lower than in any comparator having been piloted for 8 years and fully implemented countries (Figure 42). However, it should be noted that in 2015, challenges with PB remain. First, linkages the shortfall in higher education expenditure may have between the BSP, ESP, and Annual Operating Plan partly been mitigated by the additional funding accessed (AOP) are limited, and some misalignments have been through tuition fees and not accounted for in the budget identified (RGC, 2015). Second, the amounts requested (RGC, 2015). Nevertheless, the financial autonomy of in the BSP by budget entities (bottom-up approach) are higher education institutions remains limited. In the often not realistic or in concordance with previous years, context of ASEAN integration, the stark difference in making it less effective as a tool (World Bank, 2016). funding between Cambodia and neighboring countries Third, the PB structure does not adequately capture may leave Cambodia at a disadvantage in the medium program cost including capital expenditure, staff, and to long term as it tries to move up the value chain and other costs, as nearly all personnel expenditure is lumped diversify beyond garments (ADB, 2015). in one single program rather than being distributed to 74 Cambodia: A Public Expenditure Review pertinent subprograms. In addition, DP financing is not School-Based Management is a system in which fiscal included in the BSP/PB, and the categories under which decision making is delegated to school management, but it is classified in the CDC database do not correspond with parents and community stakeholders given formal with the programs and sub-programs defined. While influence and authority to participate in management DP funding is included in the AOPs, non-inclusion in decisions and monitor expenditure and activities. By planning documents hampers assessment of sub- introducing community members and parents in budget sector allocations and performance. decisions, the system attempts to induce parents to actively engage in their children’s education, both at Despite improvements, delays in the flow of funding school and at home, while providing a cost-effective to schools persist. Initially, the implementation of and well-informed check on school administrators. The program budgeting resulted in a falling execution MoEYS aims to introduce this system to complement rate for non-wage expenses (falling from roughly 90 the planned expansion of funding provided through percent during 2010-2013 to 83 percent in 2015). Two flexible SIF transfers. If implemented well, the system main reasons have been identified. First, funding was of disbursement and management will result in effective initially attached to specific sub-accounts, and budget spending decisions based on local knowledge and managers did not have the flexibility and authority to therefore effectively address school-specific needs and reallocate unspent budget from one line to another challenges. In order for the system to work, however, when required (Ung, Oung, Tep, Kann, & Pring, 2016, p. considerable training and follow-up is required to 17). To respond to this challenge, further flexibility was empower all involved stakeholders to fill their roles granted to budget managers in 2017, especially with effectively. regard to execution of the School Operating Budget (SOB). Second, a recent Public Expenditure Tracking Survey (World Bank, forthcoming) reveals that, while no significant leakages were identified, most schools Access and equity of spending only received their 2016 SOB budget allocations in March (rather than January). While there has been As enrollment increases, public spending some streamlining of internal MoEYS procedures— in education is becoming more pro-poor notably the introduction of direct bank transfers which has reduced disbursement time—significant delays Despite significant progress, Cambodia still has persist at the Provincial Departments of Education in an enrollment challenge, especially for secondary consolidating, reviewing, and processing fund requests education. After successfully having been increased and to the Provincial Treasuries (World Bank, forthcoming). 50 maintained at high levels over the past decade, primary school enrollment seems to be facing headwinds. The To promote high-quality spending decisions and primary net enrollment rate peaked at 96.4 percent accountability, increased fiscal autonomy for schools in 2011 but is estimated to have contracted to 93.9 would require finance training for administrators as percent in the 2015-16 school year. Lower secondary well as a School-Based Management system which net enrollment has increased considerably over the last involves communities in school decision making. five years, from 35 percent in 2010 to an estimated 38.9 50  Funding for SOB, which is used for basic school expenditures (excluding personnel costs and major procurement such as curriculum material), is delivered as cash advances in four installments at the beginning of each quarter, from respective provincial treasuries. It is deposited directly into each school’s bank account starting from 2015, as indicated in the inter-ministerial prakas no 366, between Ministry of Economy and Finance (MEF) and Ministry of Education, Youth and Sport (MoEYS) dated 06 April 2015. Before funds are released to schools, each Provincial Department of Educa- tion (PDE) prepares a consolidated cash advance request for all schools based on school development plans and approved budgets within respective provinces. The request is submitted to the Provincial Department of Economy and Finance (PDEF) for review and approval at the beginning of each quarter. Upon review and approval, the PDEF further submits the request for cash advance to the Provincial Treasuries (PTs), which disburses SOB into the bank accounts of each school (Ung, Oung, Tep, Kann, & Pring, 2016). Cambodia: A Public Expenditure Review 75 Table 22. Education expenditure is still pro-rich but seems to be becoming more pro-poor Share of education expenditure received by quintiles 2004 2014 Quintile Pre- Primary Secondary Tertiary Total Pre- Primary Secondary Tertiary Total primary primary 1 11% 17% 7% 3% 12% 19% 18% 16% 13% 17% 2 12% 19% 12% 6% 16% 15% 19% 18% 15% 19% 3 9% 21% 16% 12% 18% 18% 21% 20% 19% 20% 4 26% 21% 25% 19% 23% 26% 21% 22% 23% 22% 5 42% 22% 40% 60% 32% 22% 21% 23% 31% 22% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% Source: MoEYS budget 2015, EMIS enrollment data 2015, CSES 2004 and 2014. Note: Per child unit cost is based on MoEYS budget expenditure by education level from 2015 and EMIS enrollment data by education level, hence unit costs only vary by education level. This unit cost is applied to the enrollment data from CSES 2004 and 2014, hence unit cost does not vary across time. The table is an extension of the work in Phay and Tong (2014) and differs in that the analysis relies on all households (as opposed to only households with children) and that student unit costs vary by education level. percent in 2015, with gross enrollment at 53.8 percent.51 Wealthier households are more likely to enroll Overall upper secondary enrollment remains low, with a at all levels. For instance, 6, 77, and 53 percent of gross enrollment rate of 24.3 percent in 2015. Divided households in the first quintile are enrolled in pre-school, by gender, primary female net enrollment stood at 94.7 primary and secondary, respectively, compared to 7, 88, percent in 2015-2016, with gross lower secondary and and 72 percent in the fifth quintile for same levels. Non- upper secondary estimated at 56.7 and 25.2 percent, enrollment in primary and lower secondary is especially respectively—all above overall enrollment rates. concentrated among the poorest, while enrollment in upper secondary and preprimary is also low for middle- Education expenditure on pre-primary, primary, income families. Pre-primary enrollment is notably less and secondary has been pro-rich, but seems to be dependent on wealth. becoming more pro-poor. A benefit incidence analysis shows that in 2004, about 32 percent of government Income variation can explain the rural/urban variation educational expenditures on pre-primary through in enrollment in primary and lower secondary, while tertiary went to the richest quintile, while the first quintile other additional factors are at play in pre-primary received only 12 percent. The skewed benefits were and upper secondary. Enrollment at all levels is higher driven by pre-primary, secondary, and tertiary education in urban areas. Regression analysis reveals that, at the having more enrollment among wealthier households. primary and lower secondary levels, a child with the However, with increased enrollment for all quintiles, same income level in urban and rural areas is equally expenditures were more equally distributed in 2014, likely to enroll. Wealth differences explain the entire with 17 percent of expenditures going to the poorest urban enrollment gap shown in Table 23. For the pre- quintile and 22 percent going to the richest quintile primary and upper secondary levels, however, roughly of households (Table 22). By many standards this is half of the lower enrollment of children in rural areas not alarming, but it does highlight that the challenge is due to income differences, while other factors (e.g., of achieving equal enrollment for all is a challenge of lack of schools) also play a significant role. Unevenness increasing enrollment among the poor. in enrollment (across quintiles and between rural and 51  Enrolment rate data from MoEYS, EMIS department, Education Statistics. 76 Cambodia: A Public Expenditure Review Table 23. At all levels of education, non-enrollment is higher in rural areas Non-enrollment by age Preprimary Primary Lower secondary Upper Secondary 4-6 years 7-11 years 12-14 years 15-17 years Urban 53% 5% 7% 27% Rural 65% 8% 14% 49% Source: CSES 2014. Table 24. The reasons for not enrolling in school vary by age group and location Preprimary Primary Secondary age 1 4-6 years 7-11 years 2-17 years All NE provinces All NE provinces All NE provinces Don't want to 4 7 31 40 19 32 Did not do well in 0 0 7 3 14 11 school Lack of school/ 1 5 9 28 4 11 teacher Financial reasons 2 2 20 19 58 45 Disability/illness 0 0 5 0 2 1 Too young 92 87 26 10 1 0 Other 0 0 2 0 1 0 Total 100 100 100 100 100 100 Source: CSES 2014. Figure 43. Non-enrollment in the Northeastern provinces is partly explained by long travel distances Distance to school and non-enrollment, by province 3,000 2,500 Average distance from villages Ratanak Kiri to upper secondary school Stung Treng 2,000 Mondul Kiri 1,500 Koh Kong Preah Sihanouk Preah Vihear Kratie Siemreap Prey Veng Oddar Meanchey 1,000 Kampong Thom Pailin Battambang Kampong Chhnang 500 Kampot Pursat Banteay Meanchey Svay Rieng Kampong Cham Phnom Penh Kandal Kampong Speu 0 Takeo 1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Share of children aged 13-18 not in school, 2015 Note: Data is GPS location of villages and secondary schools, and out-of-school population is from CSES 2014. Distance is based on GPS locations of upper secondary schools and villages and based on nearest neighbor matching of distance measured “as the crow flies.” Source: EMIS from MoEYS (2015). Cambodia: A Public Expenditure Review 77 urban areas) is common for most developing countries, measures between villages and the local enrollment with conditional cash transfers and scholarship schemes levels show the same pattern, as large shares of youth being proven mitigating measures (Evans & Popova, are not attending secondary school in provinces where 2016). Pilot scholarship projects in Cambodia have villages are situated farther away from schools (Figure obtained strong results in retention/enrollment, although 43). Increasing enrollment efficiently thus requires they do not seem to have resulted in improved learning localized solutions. outcomes (World Bank, 2012; ADB, 2012). Following multiple pilots funded by DPs, the RGC has assumed responsibility for and included a scholarship component in its budget. The planned expansion of this scheme is Expenditure efficiency expected to have a positive impact on enrollment. Cambodia has a large net shortage of Most families of children who are out of school in teachers in pre-primary and primary, pre-primary and primary are reporting reasons exacerbated by poor distribution of other than financial and access issues. Out of 100 teachers families with children aged 4-6 who are out of school, 92 percent of household heads give age (being too young) Cambodia’s large net shortage of teachers in pre- as the main reason. In the case of primary education, primary and primary translates into High student- 31 percent report that their children aged 7-11 do not teacher ratios compared to other countries. Based want to go to school, and 26 percent think they are too on class-size norms for 2014, at the national level, young (Table 24). In contrast, 58 percent of households Cambodia had a net shortage of about 8,000 pre- with children aged 12-17 who are not enrolled in primary teachers and 22,000 primary teachers.52 For secondary education point to financial reasons, since primary, there is a downwards trend in student-teacher the opportunity cost of being at school (instead of ratios, as the student population appears to be falling. working) is perceived as high. For pre-primary, booming enrollment is pushing student- teacher ratios higher. In both cases, the student-teacher In some remote provinces, lack of schools and ratios are significantly above the averages of any other teachers seems to be a significant barrier to comparator country (Figure 44). enrollment. As presented in Table 24 above, the percentage of households who cite lack of schools The net shortages mask a more complex problem, and teachers as the main reason for non-enrollment as most provinces have both schools with shortages is significantly higher for provinces in the Northeast and schools with a surplus of teachers. A series of of the country. In particular, for upper secondary, rules allowing teachers to relocate after a few years of household wealth does not explain all non-enrollment service plus lack of enforcement of teacher-to-students for provinces such as Stung Treng, Oddar Meanchey, and teacher-to-classroom norms have resulted in a and Ratanak Kiri, with access challenges at play. At significant number of teachers relocating to schools with other education levels, only Ratanak Kiri consistently surplus. While Cambodia has a total shortage of 39,000 has lower enrollment levels that cannot be explained by teachers, or more than 44 percent of the current number household wealth. Households’ own reporting of access of teachers, it also has around 8,000 surplus teachers barriers as reasons for not enrolling is consistent with in schools with redundancies (those which would be spatial analysis, as secondary and upper secondary complying with the norm even if those teachers were not schools are scanty in the Northeast provinces. Distance there) (Table 25). The problem is particularly noticeable 52  The calculation of teacher shortages and surpluses is based on the 2014 student teacher and class norms. 78 Cambodia: A Public Expenditure Review Figure 44. Cambodia has high student-teacher ratios compared to other countries 50 45 40 35 30 25 20 15 10 5 0 Cambodia Bangladesh Lao PDR Thailand Vietnam Low-income Lower Middle-income middle-income Pre-primary* Primary Lower secondary Upper secondary* Note: * indicates old data included. The pre-primary number for Bangladesh is from 2002, the upper secondary number for Cambodia is from 2007. Source: UNSECO Institute of Statistics (UIS), latest year available. Table 25. Cambodia has both schools with teacher surpluses and shortages (2015) Preprimary Primary Secondary Total National Current teachers 4,537 44,884 40,924 90,345 Surplus teachers 34 2,718 5,403 8,155 Shortage teachers -8,005 -25,061 -6,789 -39,855 Net shortage teachers -7,971 -22,343 -1,386 -31,700 Urban Current teachers 957 9,712 12,636 23,305 Surplus teachers 25 1,162 2,794 3,981 Shortage teachers -1,172 -1,919 -766 -3,857 Net shortage teachers -1,147 -757 2,028 124 Rural Current teachers 3,580 35,172 28,288 67,040 Surplus teachers 9 1,556 2,609 4,174 Shortage teachers -6,833 -23,142 -6,023 -35,998 Net shortage teachers -6,824 -21,586 -3,414 -31,824 Source: CSES 2014. in secondary education, as teachers in schools with larger in rural areas as shown in Table 25, the challenge redundancies amount to 80 percent of total shortages. is not just excess teachers being located far away in the nation’s capital and remote villages having no Some of the teacher shortages could be solved teachers, although this is part of the problem. In fact, through “reasonable” relocations or incentives most provinces have both schools with surplus and for teachers to teach in neighboring schools with schools with shortages. A large share of the shortages shortages. Notably, while shortages are significantly at the primary level, in the provinces of Phnom Penh Cambodia: A Public Expenditure Review 79 and Banteay Manchey, could be addressed by simply locations with shortages. While no doubt mitigating, relocating teachers from surplus schools to deficit these measures have proven insufficient. schools. At the secondary level, Phnom Penh, Kandal, Preah Vihear, Pursat, Strung Treng, Kracheh, and Koh The number of teachers teaching in schools where Kong would still have excess teachers even if all the they are not needed increases every year. The number schools with shortages were “filled” through relocations. of teachers in schools with redundancies grew by more In total, 43 percent of primary schools and 28 percent than 1,300 between 2013 and 2015, an increase of of secondary schools with shortages are located within almost 20 percent (Figure 45). As of 2015, an estimated 5 kilometers (“as the crow flies”) of the nearest school 10 percent of teachers’ salaries were going to teachers with a surplus. who were teaching in schools with redundancies. The monetary loss is equivalent to the annual salary of The cost of misallocated teachers is more than 10,000 primary or pre-primary teachers— significant and rising this is arguably a large loss of resources in a budget- constrained environment. Authorities have undertaken a series of measures to try to address teacher shortages in the short Fulfilling all teacher norms by hiring new teachers term, but this has not been sufficient. The fact that would increase teacher salary costs by 40 percent, some teachers are located at schools where they are but this would be reduced to 30 percent if teachers not needed is not a new finding, and solutions to the were distributed optimally. If authorities were to problem have been discussed for a while. To partly address the entirety of teacher shortages by hiring address this challenge, contract teachers, receiving new teachers, this would drive the MoEYS payroll up a higher salary, have been introduced in hard-to-fill by 40 percent. Alternatively, if all teachers teaching in positions (often in remote areas). Monetary bonuses for schools with redundancies were relocated, this could remote schools have also been introduced to increase result in savings equivalent to 10 percent of the payroll. incentives for teachers to take less attractive positions, Considering the continuing flow of teachers from deficit and efforts are being made to recruit new teachers in to surplus schools currently, it is likely that the loss (and Figure 45. The number of excess teachers has been growing each year 6,000 Number of excess teachers in: 5,000 4,000 3,000 2,000 1,000 0 2013 2014 2015 Pre-primary Primary Secondary Source: EMIS 2013-2015 and World Bank staff calculation. 80 Cambodia: A Public Expenditure Review consequently, potential gains from relocation) would per student, both in 2015 and as an average over the grow beyond these magnitudes if unaddressed. period 2010-2015. Similarly, provinces such as Stung Treng, Kampot, and Banteay Meanchey consistently Variation in provincial student unit costs had among the highest government expenditure per are driven by challenges in service student, both in 2015 and over the 2010-15 period. delivery, especially to remote areas It seems that variation in per cost-equivalent student Per student funding varies significantly across spending is driven not only by teacher shortages but provinces but is not targeted in favor of the poorer also by the difficulty of service delivery (remoteness). provinces. Figure 46 shows that per cost-equivalent It could be thought that variation in per cost-equivalent student funding53 varies significantly across provinces, student spending would be determined entirely by with some provinces spending up to two times more per existing teacher shortages. However, this does not cost-equivalent student than others. However, variations seem to be the case, as regional student unit cost in both DP and government funding per student do not displays a similar ranking, even when corrected for net seem to be associated with poverty incidence across cost of the misallocated and missing teachers. Regional provinces. While there is some variation over time, student unit costs are highly correlated with population provinces like Kampong Speu, Siem Reap, and Prey density and average school size, which indicates that Veng had among the lowest government expenditures variation in unit costs are at least partly driven by the Figure 46. Provincial variations in expenditures per student are not associated with poverty incidence 900 35 800 Expenditure per student, thousand riel 30 700 25 600 Poverty headcount 500 20 400 15 300 10 200 5 100 0 0 Preah Vihear Phnom Penh Tboung Khmum Preah Sihanouk Kandal Svay Rieng Kep Kampot Koh Kong Takeo Prey Veng Kampong Cham Banteay Meanchey Pailin Siem Reap Kampong Speu Kampong Chhnang Battambang Mondul Kiri Ratanak Kiri Pursat Kampong Thom Kracheh Stung Treng Otdar Meanchey DP expenditure Government expenditure Share poor in Province Source: World Bank staff calculations based on province-level education expenditures, EMIS student data for 2015 and MoP (2014). 53  Per student funding is calculated as all expenditures on preprimary, primary, and secondary in each province divided by number of “primary equivalent students.” A secondary student cost two times more than a primary student on average, and a pre-primary student costs 1.5 times more than a primary student on average. The total number of primary equivalent students is therefore the number of primary students plus two times the number of secondary students plus 1.5 times the number of pre-primary students in each province. Utilizing primary equivalent students ensures that provinces with higher pre-primary and secondary enrollment are comparable to other provinces, avoiding the appearance of higher student unit costs just because of higher enrollment shares of costlier students. Cambodia: A Public Expenditure Review 81 Figure 47. Regional student unit costs are driven by numerous factors Drivers of regional student unit costs, in thousand riel (2015) 700 700 Stung Treng Stung Treng Ratanak Kiri Ratanak Kiri Siem Reap 600 600 Koh Kong Pailin Koh Kong Pailin Battambang Siem Reap Prey Veng Preah Vihear Preah Vihear Prey Veng Otdar Meanchey Kep Kep Otdar Meanchey Battambang Svay Rieng 500 Svay Rieng Mondul Kiri 500 Mondul Kiri Pursat Kampong Chhnang Pursat Kampong Cham Kampong Chhnang Kampong Cham Takeo Kampong Thom Kracheh 400 Takeo Kampong Speu 400 Preah Sihanouk Kandal Kracheh Kandal Kampot Kampot Tboung Khmum Preah Sihanouk Phnom Penh Tboung Khmum Kampong Thom Banteay Meanchey 300 300 Kampong Speu Phnom Penh Banteay Meanchey 0% 2% 4% 6% 8% 10% 0 100000 200000 300000 Percentage of all employees working at province and regional level Number of cost equivalent students 700 700 Stung Treng Stung Treng Ratanak Kiri Ratanak Kiri 600 600 Koh Kong Koh Kong Pailin Pailin Preah Vihear Preah Vihear Otdar Meanchey Otdar Meanchey Kep Kep Prey Veng 500 Battambang Prey Veng 500 Mondul Kiri Siem Reap Mondul Kiri Battambang Pursat Svay Rieng Pursat Kampong Cham Kampong Thom Takeo Kampong Thom Takeo Kracheh Kampong Speu 400 Kandal 400 Kracheh Kandal Kampong Chhnang Kampong Cham Siem Reap Tboung Khmum Kampong Speu Svay Rieng Phnom Penh Kampot Preah Sihanouk Kampot 300 Preah Sihanouk 300 Kampong Chhnang Phnom Penh Banteay Meanchey Banteay Meanchey 2 4 6 8 5 5.5 6 6.5 7 Log population density Log average school size Source: World Bank staff calculations based on EMIS and provisional education expenditures. difficulty of providing educational services, with a more KHR 500,000 have at least 5 percent of the workforce scattered population being costlier to serve (Figure 47). employed in administrative functions at province and Notably, factors such as the share of the population regional level. living in urban areas and the share of students in two shifts are not correlated with student unit costs. Additional increases in enrollment are likely to come at higher student unit costs. Analysis reveals that the Larger regions also seem to benefit from economy of regions with lower shares of children enrolled are also the scale in administration costs. Regions with a relatively regions with higher average unit costs. This is consistent small enrolled population have higher unit costs, which with the abovementioned finding that these regions in part are driven by higher ratios of administrative have lower population density and lower average school personnel to teachers/students (Figure 47). Notably, all sizes. Hence, reaching the remaining children is very regions with unit costs under KHR 500,000 per student likely to be more difficult and more expensive compared have less than 5 percent of the workforce employed in to the students already enrolled. Future enrollment administrative functions at the provincial and regional costs per student are therefore likely to increase. There levels, while those regions with unit costs higher than is evidence that small secondary schools in scarcely 82 Cambodia: A Public Expenditure Review populated areas already have their students traveling of students) was KHR 20,000 for pre-primary, KHR the farthest, and most of the additional enrollment 14,000 for primary, and KHR 25,000 for secondary in would need to take place in these schools (and would 2015. As a recent PETS survey indicated, however, the be costlier). The additional salary bonuses to teachers in amount received by small schools could frequently be remote locations also raise costs. However, the higher too low to meet minimum spending requirements (World costs could be partly offset if the number of positions Bank, forthcoming). at the provincial and regional levels do not increase at the same rate as additional enrollment. The variation in Evidence also indicates that the School Improvement administrative cost across regions would decrease, as Grants (SIG) are insufficient to ensure that small regions with relatively higher administrative costs also schools are in good condition. While the SOB is tend to have larger scope for additional enrollment. government funding covering operating expenses, the SIG scheme is DP funded and finances capital Greater utilization of per-student funding expenditure at the school level. Small schools receive a could have many advantages larger SIG subsidy per student but also seem to be in worse physical shape, as assessed using a school quality The School Operating Budget (SOB) is a relatively index, 58 which would mean funding may be insufficient to small amount that does not meet school needs. The cover needs. This is especially the case for primary and SOB is a direct transfer of government funding to schools, secondary schools with less than 100 students, which covering operating expenditures, but excluding salary are in much worse condition than any school with over payments and major procurement. Only 3.3 percent 54 55 200 students.59 Ensuring sufficient allocations will be of government educational expenditures on pre-primary, essential as preparations are underway for a government primary, and secondary education are distributed takeover of the SIG component and the merger of the through the SOB formula.56 It is equivalent to less than SOB and SIG systems into a School Improvement Fund 20 percent of all non-wage expenditures (21 percent at (SIF) system, fully financed by the government. the pre-primary and primary levels and 17 percent at the secondary level), an amount that is viewed as insufficient There is evidence that providing funding directly for general operation of schools (ITAD, 2016). The SOB to schools through SOB and SIG is contributing formula assumes different degrees of economy of scale to improved facilities and equipment. International in pre-primary, primary and secondary, which results experience with money following the student highlights in different net per-student funding to schools in pre- advantages such as greater transparency, greater primary, primary, and secondary as a function of school predictability, and a school system more adaptable size. At the median, the total net per student funding 57 to change in enrollment patterns (Alonso & Sánchez, (fixed amount plus variable amount, divided by number 2011). Evidence indicates that combined SOB and 54  The SOB finances expenditures like office supplies, small scale repair of school equipment, books for library, arts and crafts, clean water, electricity etc. 55  Estimated by multiplying school characteristics in EMIS 2015 with SOB subsidies divided by MoEYS combined budget for pre-primary, primary, and secondary. There is a small difference in using EMIS data times the formula which amounts SOB to 45.0 million riels, compared to MoEYS budget number of 45.6 million riels in 2015. 56  The SOB is based on a funding formula mixing a fixed amount per school and a variable amount per student. Both components vary with the number of classrooms in the school. In addition, a total of 300 schools (pre-primary, primary, and secondary) are designated as disadvantaged schools, and these schools receive higher amounts of both subsidies. 57  For primary, for instance, the formula is largely a per-student subsidy only, as the fixed amount is relatively small and only matters for very small schools with less than 50 primary students (and only 4 percent of primary schools have less than 50 students). In contrast, 36 percent of secondary schools and 93 percent of pre-primary schools have less than 100 students, and these experience significant boosts in net subsidy per student. 58  The index of school quality is the first component of a principle component analysis based on presence of water, toilets, black board, chairs, desk, playground, quality of structures (concrete and wood). 59  While per-student funding may be higher for small schools, it is difficult to see how significant improvements in school facilities could be made with a SIG disbursement of USD 118 (the average received by the lowest quintile of stand-alone primary schools in 2016) or USD 196 (for the cor- responding quintile for lower secondary schools). Combined SOB and SIG allocations or operational funds were disbursed as low as USD 248, with the lowest quintile receiving between USD 248 and USD 1,092. In recognition of this, the SOB allocations for 2017 included a “top-up” of KHR 1 million per school as a pilot allocation increase. Cambodia: A Public Expenditure Review 83 SIG funds are at the moment positively correlated with school quality (e.g., learning materials, library, furniture), Quality and learning outcomes financial quality (compliance with ministerial financial National student test scores reveal significant management guidelines), and environmental quality differences between urban and rural areas, (safety, utilities) (World Bank, forthcoming). Providing between boys and girls, and across provinces. On funding directly to schools, combined with more average, students located in urban areas answer 55 flexibility in its use at the school level, could lead to percent of questions correctly, compared to less than greater efficiency as money could be directed toward 50 percent in rural areas. This is a significant difference, areas where the school sees a larger need. but Figure 48 also shows that the 50 percent of urban students in the 25- 75th percentile score between In addition, a higher per student component, 42 and 68 (illustrated by the solid box), while the 50 combined with improved supervision and percent of rural students in the same percentiles score accountability, could provide further motivation between 36 and 52. Hence, there is also a substantial for schools to maximize enrollment and student variation within the different groups, and overlap in presence. Analysis in Section 4.3 indicates that scores. Similar patterns can also be seen for provinces some primary students might not enroll due to lack of and male and female students, with female students knowledge or motivation, as opposed to financial cost having statistically significantly higher scores than or access to schools, while there is evidence of better male students (average difference of 2.5 percentage outcomes for students that have higher attendance points). Decomposing student outcomes into variation (Section 4.5). As such, implementation of per-student within and between schools shows that about half the funding could potentially result in an additional variation can be explained by variation within the same improvement in both enrolment and student outcomes class. through incentive externalities. Figure 48. Significant differences in test scores can be seen between males and females and between urban and rural students Share of correct answers in student tests 100 100 80 80 60 60 40 40 20 20 0 0 Female Male Rural Urban Source: MoEYS Quality Assurance Department test scores for 3rd and 8th grade. 84 Cambodia: A Public Expenditure Review Table 26. Regressions for 3rd and 8th grade test scores Both grades 3rd grade 8th grade coef se coef se coef se Student Characteristics Student social economic characteristics 1.5*** 0.3 0.5 0.4 1.9*** 0.3 Student absenteeism -1.5*** 0.2 -0.6** 0.3 -1.8*** 0.2 Student's homework 0.7** 0.3 0.8 0.6 0.6** 0.2 Teacher Characteristics Teacher received on job training -0.3 1.0 -6.4*** 2.3 1.0 1.0 2 years of teaching training -1.7 1.6 -1.0 1.6 6.7*** 1.8 3 years of teaching training -0.5 2.1 -1.2 4.4 8.9*** 2.1 4 years of teaching training 2.7 3.1 13.1*** 3.0 Standardized teaching behavior in class room 2.5*** 0.3 1.9*** 0.5 2.4*** 0.3 Teacher understand curriculum 1.5* 0.9 3.6*** 1.3 0.7 1.0 Teachers absenteeism 0.1 0.0 0.0 0.1 0.1** 0.1 Hours spend on preparation 0.0 0.0 -0.0 0.0 0.0 0.0 School Characteristics Education of School Principal 0.3 0.3 1.1** 0.5 -0.0 0.3 Total Service of School Principal 0.0 0.1 -0.1 0.1 0.0 0.1 School size 0.0*** 0.0 0.0*** 0.0 0.0** 0.0 Schools size squared -0.0*** 0.0 -0.0*** 0.0 -0.0 0.0 Student teacher ratio -0.0 0.0 0.0 0.0 -0.1** 0.0 School has more than one shift -3.2*** 1.0 -1.4 1.6 -2.6*** 1.0 Distance from school to commune: Kilometer -0.1 0.1 0.1 0.2 -0.4** 0.2 Urban school 3.2*** 1.2 4.1* 2.3 0.4 1.3 Fixed control variables Grade 8 dummy 6.9*** 1.5 Math test dummy -2.9*** 0.9 5.9*** 1.2 -11.8*** 0.9 Physics test dummy 1.7 1.1 -2.7*** 1.0 Constant 38.9*** 3.2 27.8*** 4.9 49.8*** 3.1 Number of observations 21,388 8,590 12,798 Adjusted R2 0.186 0.205 0.170 Source: Authors’ calculations based on EQUAD 2014 and 2015. note: .01 - ***; .05 - **; .1 - *; Standard errors take school level clustering into account. Additional regressions and robustness test are available in (Sohnesen, 2016). Cambodia: A Public Expenditure Review 85 Teacher characteristics and behavior in Muralidharan (2012) further found that an individual classroom significantly affect student performance pay program could be 15 to 20 times performance more cost-effective than traditional measures (such as increasing the number of teachers) in improving Regression results 60 show that teacher learning outcomes. characteristics affect student performance more than socioeconomic background. For students Further, there is notable scope for increasing in eighth grade, a teacher with four years or more learning outcomes through improved in-classroom of teacher training is associated with an increase of teaching in Cambodia. This is reflected in Benveniste, 13.1 percentage points in student scores (Table 26). Marshall, and Aranjo (2008) and Tandon and Fukao Teacher behavior in the classroom has an impact on 61 (2015), with the latter listing encouragement of stronger both eighth grade (+2.4 points) and third grade (+1.9) classroom performance as one of their three key policy students, the levels for which test data was available. pillars for improvement of learning outcomes. A high For third grade, even when on-the-job training does not degree of variation can be seen in the in-classroom seem to have a positive impact, teacher understanding behavior index: for example, only 1 percent of teachers of the curriculum does have a significant effect on in the bottom (first) quintile of the index (the teachers student scores (+3.6). with the worst behavior) correct all homework, while 93 percent of teachers always correct all the homework in There is clear consensus from studies on learning the top quintile. Similarly, 6 percent of teachers in the outcomes that teachers’ effort, as well as first quintile have students at the blackboard and 7 understanding of material and teaching methods, percent use multiple choice tests in their teaching, while are essential to results. In a review of the existing 75 and 46 percent of teachers in the fifth quintile have literature and assessments on learning outcomes in students at the blackboard and use multiple choice developing countries, Evans and Popova (2016) found tests. The significant positive relationship between that, while there was large variation in results and teacher behavior and outcomes is a robust result found recommendations, consensus on effectiveness was in both 3rd and 8th grade, in urban and rural areas, and evident for the introduction of high-quality pedagogical for all subjects (Sohnesen, 2016). interventions (curriculum, teaching methods, and technological aids) and matching of teaching to School characteristics and students’ own student learning. However, this depends critically on efforts also matter teachers’ understanding of the material, technological aids, and pedagogic teaching methods. In the Students in large schools perform better. School size absence of teacher training and continuous follow- is significantly correlated with student scores, with large up, the results of curriculum improvements and other schools obtaining higher scores. Larger schools have measures are much less positive. Further studies , 62 better inputs both in terms of education of teachers indicate that teacher effort may matter more than and physical inputs, but several other factors could also qualifications, with teacher performance pay schemes explain why larger schools perform better, including that are connected to student test scores or teacher better management and better utilization of teachers absenteeism achieving significant positive results. (e.g., through specialization). Interestingly, the student- 60  The regression corrects for aspects of the students that are beyond government control, such as education and employment of parents. All these aspects are controlled for with a Social Economic Characteristics (SEC) index.. 61  Teacher behavior in classroom is captured by eight questions answered by students on teachers’ teaching styles, including if teachers are angry with students, help them with homework, and allow them to come to the blackboard. 62  See Glewwe & Muralidharan (2016), Duflo, Hanna & Ryan (2012), Contreras & Rau (2012) and Muralidharan (2012) 86 Cambodia: A Public Expenditure Review teacher ratio alone does not seem to have a significant covered by DPs (as has been done with the scholarship impact on student test scores. School size and excess/ programs, for example). In this regard, it is worth noting surplus of teachers do not seem to be correlated, while that the 2017 budget included a significant increase in students attending a school with more than one shift government-funded capital allocation, with a planned obtain lower scores. capital allocation of KHR 90 billion riel, up from zero in 2014, 35 billion in 2015, and 40 billion in 2016. While still Students’ own efforts matter and could be not a large allocation, this signals that the government supported by the active engagement of parents. is willing to take more responsibility by scaling up its Students who are more absent have lower scores, activity. If Cambodia is to achieve the ESP objective of and students who do more homework have higher adequately funded education, this trend will need to scores in 8th grade. This is true even among students continue. with similar socioeconomic characteristics, teacher characteristics, and school characteristics. The 6 To scale up stewardship of the sector, continued percent of 3rd grade students and 9 percent of 8th improvements in the allocation and use of resources grade students reporting more than 11 missed days are needed through stronger program budgeting. of school have scores that are on average 7 and 11 While the ESP provides the direction for the education percent lower, respectively, than students with no sector, the linkages between policy and funding are still absence. Marshall et al. (2012) found a similarly large 63 weak. Addressing this challenge would require a series impact of student absences on student scores in their of measures, including (i) improving clarity of allocations analysis of 6th graders in 2006 and 2009. Although to different programs by unwrapping lumped wages not an input the government has direct control over, and mapping personnel costs by program in budgets it might be an aspect that could be influenced at low and accounting reports; (ii) moving toward more cost through information campaigning or other forms realistic BSPs by facilitating information exchange of motivation for students, parents, teachers, and between MEF and MoEYS (including on the available schools as discussed in the following section. budget envelope); (iii) continuing to provide change management and leadership training for managers and budget entities in charge of implementation; (iv) shifting responsibility and accountability to budget managers Policy options by allowing all expenditure receipts to be retained at the spending ministry for audit at a later date, in line Government expenditure stewardship with the government budget reform strategy; and (v) could be strengthened further improving DP coordination mechanisms and including funded projects in the BSP, since they will The government should keep expanding the funding have an impact on the results obtained. The Annual for, and ownership of, education policies. Greater Operational Plans already include DP funding in a quite stewardship of the sector could be taken, as DPs detailed manner. If a similar—albeit tentative—inclusion currently fund 69 percent of all non-wage expenditures, could be made in the BSP, plans and outcomes leaving most policy-relevant expenditures to DPs. Since could be compared with DP contributions taken into DP funding to education as a percentage of GDP is account. The above is in line with the ESP, which expected to remain flat and eventually start declining, sets out objectives of continued improvement and the government should expect to keep assimilating and clarification of BSP processes, while also emphasizing funding an increasing number of functions previously management capacity improvement. 63  Absence can also be correlated with ability and/or interest, but the analysis is unable to control for this directly. Cambodia: A Public Expenditure Review 87 Increasing both funding and accountability at the education spending in Cambodia is by far the lowest school level could play a key role in improving among comparator countries, at 5 percent (highest schools’ inputs and outcomes going forward. A estimate) of total sector spending. While ensuring recent report (ITAD, 2016) details the potential benefits access to lower levels of education, attention should and challenges of combining SOB and SIG funding also be given to higher education programs in line into a single school funding program, including moving with Cambodia’s IDP (RGC, 2015b) to ensure that the toward medium-term funding, which would allow necessary skills are available at a sufficient quality to improved school financial planning. In 2017, authorities meet national strategic goals. The need for advanced already moved to reduce SOB spending restrictions skills warrants increased expenditure in higher and were preparing for the merger of the SOB and SIG education, with a focus on quality improvement. The systems into a SIF system as mentioned earlier. This ESP recognizes this and targets increased spending would be a good opportunity to (i) reduce the number and improved quality in higher education. In general, of annual transfers and simplify auditing standards greater financial autonomy could allow the sub-sector and procedures; (ii) grant further flexibility with regard to derive more benefits from tuition fees than under to how funds are spent (instead of earmarking), while current arrangements. This would provide funding reinforcing accountability mechanisms; (iii) include a for advancing their research facilities and human base amount for travel allowances to allow schools resources to implement an updated curriculum. In to cover immediate teacher shortages (particularly return for increased autonomy at the institutional level, for subject teachers) by offering travel compensation the management of public higher education institutions for teachers in nearby schools with spare capacity; could be held legally accountable through accreditation (iv) revise the formula for the distribution of funding, and quality assurance mechanisms. The block grant as while there is currently some allowance for introduced by the MoEYS is a step in the right direction, “disadvantaged schools,” the formula does not seem but there is still potential for substantial expansion of to take specific school characteristics into account autonomy. sufficiently. An assessment of general school needs based on size, location, and education level should Cost-efficient measures could be used to guide future adjustments to the SOB formula (now SIF) expand enrollment so each school receives sufficient funding to cover its operational needs. If managed well, a gradual increase Future increases in enrollment are likely to be in the SIF allocation is likely to allow localized solutions costlier and will require interventions that are and improved spending execution and efficiency. These specific to the challenges at different school levels measures could contribute significantly towards the and geographical locations. Additional enrollment and ESP target of improved school quality and improved expansion of the school network is likely to be in more disbursement procedures. remote areas and areas with lower student density. As discussed earlier, serving these areas is more In the case of higher education, which is currently expensive than serving the existing school network. underfunded in Cambodia, increased financial In the Northeast and other remote areas of Cambodia autonomy combined with accreditation and quality where physical access is still a challenge to enrollment, assurance mechanisms are important priorities. local cost-benefit analysis should be used to determine Tertiary education spending is usually not pro-poor and whether expanding access through construction of may not be viewed as the top priority in developing more schools, reducing access costs through transport, economies that are focusing on increasing primary or utilizing housing or grant options are optimal solutions and secondary enrollment. However, estimated higher in these provinces. 88 Cambodia: A Public Expenditure Review At the pre-primary and primary levels, awareness This measure would support the IDP, while being in line and communication campaigns targeting both with the ESP policy objective of increasing enrolment parents and children would be needed to enroll for disadvantaged students through income- and merit- those out of school. Spreading knowledge and based scholarships. changing expectations might be a cost-effective intervention for increasing enrollment further, as many Improved allocation of resources would school age-children are not enrolled because parents help Increase expenditure efficiency perceive them as being too young as discussed earlier. This would require targeted interventions such Some teacher shortages could be solved through as awareness campaigns. Nonetheless, it is worth “reasonable” reallocations. In 2015, about 10 percent noting that this measure may not be effective if not of MoEYS salaries were going to teachers teaching in accompanied by construction of new preschools, schools where they are not strictly needed to maintain where there is currently a shortage. The ESP targets student-teacher norms (redundant). About half of the increased ECE enrolment and provision and quality teacher shortage could be solved if all excess teachers improvement in both primary and ECE, with reference were relocated to schools with shortages. Some of to parent education on benefits of ECE, indicating these relocations could be done with minimal cost intent to make efforts on both fronts. to individuals. As noted earlier, 43 percent of primary schools with a shortage of teachers have a school For secondary education, the expansion of grant with a surplus of teachers within 5 kilometers. While programs could help reduce the opportunity cost wholesale reallocations may be unfeasible, this calls of being at school. In the secondary-level age group, for pragmatic reallocation of teachers to nearby 58 percent of the children that are out of school cite schools, especially for primary education. When not financial reasons as the main cause. Past experiences possible, teacher shortages in certain schools could be have shown that scholarship schemes do increase mitigated by including allocations in SOBs to provide enrollment, although limited impact was found on monetary incentives for teachers in nearby schools/ learning outcomes (Evans & Popova, 2016; World Bank, locations to travel and do shifts at the school in need. 2012). Expanding the coverage and funding of current At the secondary level, this allocation could allow deficit student scholarship programs would help mitigate the schools to meet specific needs by enabling travel and direct and opportunity costs of schooling, especially compensation for subject teachers from nearby surplus for secondary education. This is recognized in the ESP, schools, potentially even allowing one teacher to cover which targets expansion of scholarship programs for multiple minor deficits. primary and secondary levels. An effective and long-term solution to poor Income- and merit-based scholarship schemes allocation of teachers would need to include teacher could be used to expand higher education management reform and enforcement of student- enrollment, especially for areas of study that are teacher ratio norms. While school teachers have a higher strategic priorities. These schemes, combined right to relocate after a set number of years of teaching, with similarly targeted education quality improvement they should not go to schools that are already above measures, could be provided for students entering IDP the teacher-to-student threshold. Yet common practice priority courses such as natural sciences and engineering is that this takes place despite existing norms. Unless to ensure the supply of desirable high-quality skills. Such the flow of teachers from deficit to surplus schools is schemes would not only advance skills development but stopped and rates of teacher absenteeism are reduced, also help maintain the pro-poorness of overall spending. they will remain a perpetual drain on government Cambodia: A Public Expenditure Review 89 resources and educational quality. While many remedial technologies to monitor teacher absenteeism would also measures have already been implemented and other be advisable. This should be considered when meeting remedial action can help mitigate the problem in ESP objectives of revising guidelines for performance the short term, improved teacher management and management, capacity assessment, career path enforcement of existing norms for student-teacher development and extensive plans for teacher training ratios and instructional hours are essential to reducing improvements. the suboptimal use of resources in the medium to long term. Hence, a first and very critical step is to start Increased parental knowledge and involvement enforcing the existing ratio norms to stop the flow of could also improve learning outcomes. Students teachers to schools where they are not needed. As that show up to school every day also perform better in done in Uganda and other countries where absenteeism school, hence involving parents to encourage students’ is prevalent, school principals and students could also full-time participation could improve learning outcomes. use mobile technologies to report teacher absence, with International experience, from Mexico among other the information being transmitted to a central database countries shows that active parental engagement can at the Ministry of Education (World Bank, 2016). The have very positive impacts on school management and above findings could be helpful in ESP policy actions student outcomes (Gertler, Patrinos, & Rubio, 2012). In including revision of teacher staffing norms, deployment a review of studies, Ganimian & Murnane (2016) found principles and financial management. that more informed and involved parents improved both retention and learning outcomes of students. Nguyen Increasing education quality is another (2009) found that teachers informing parents and key reform priority children of economic returns to schooling was effective in increasing attendance and learning outcomes, while Given that the largest spending is on salaries which outreach by other role models obtained mixed results will continue to increase, future increases should be based on personal backgrounds.64 The School Based leveraged to raise the quality of teaching. Educational Management policy foresees community and parents outcomes are significantly linked to teachers’ behavior engagement in school-related meetings, which is a in the classroom and understanding of the curriculum welcome step expected to contribute to improved and as is rightly also reflected in policy priorities in ESP. learning outcomes. These findings support the ESP Part of the quality challenge could be addressed with strategy intended to establish a quality assurance the provision of teacher training. Although both the framework involving parental participation. government and DPs allocate some resources (KHR 34 billion in 2015) to teacher training, it is done in a Notably, small schools are in need of special patchy way, as recognized by stakeholders. The current attention with regard to quality. To increase the quality Teacher Policy Action Plan is aimed at addressing of poorly performing schools, small schools need special some of these challenges, emphasizing mentoring attention as they have both the poorest inputs and programs in the short to medium term, and introducing the poorest outcomes. This is related to the fact that the requirement that all teachers must hold bachelor small schools are the ones in poorest condition (due to degrees in the long term. In addition, authorities could higher administrative costs and low absolute values of consider linking future increases in salaries to teachers’ operational funds received) as well as to the fact that the performance or passing of quality-enhancing teaching best teachers self-select themselves into larger schools certificates, based on knowledge of the curriculum and (to attain more status and extracurricular economic classroom behavior “best practices.” Introduction of opportunities). A comprehensive solution would thus 64  The method of dissemination matters. A study in Indonesia (Cerdan-Infantes & Filmer, 2015) found that informing parents through facilitated meetings or through mobile phone text messages yielded increased parental participation, while leaflets and letters were ineffective. 90 Cambodia: A Public Expenditure Review require the implementation of a range of policy options presented above to ensure sufficient operational funds as well as to improve cost efficiency and teaching quality, including the enforcement of existing rules on relocation and the increase in funding directed to schools. For the proposed reforms to succeed, administration and management will also need to be strengthened The increased focus on School-Based Management is expected to help increase enrollment and improve learning outcomes, provided that it is accompanied by some additional elements. For example, the recent introduction by authorities of increased flexibility in the use of funding at the school level and the planned increase in funding channeled through the new SIF will require a faster and well-functioning authorization and disbursement system. Improved financial recordkeeping and collection of EMIS indicators will also be necessary for more efficient public spending and increased enrollment / reduced absenteeism. Improving administrative systems to ensure the quality of classroom interactions, through teacher allocation, training and certification, and accountability, will also be crucial to ensure the provision of quality education for all Cambodians. The ESP outlines ambitious objectives of improvement in planning, disbursement, monitoring and reporting, thus recognizing these challenges. Cambodia: A Public Expenditure Review 91 Table 27. Policy options to improve expenditure efficiency and outcomes in education Challenge Short-term policy options Medium-term policy options (1-2 years) (3+ years) Two-thirds of non-wage spending Introduce further delegation of Complete progress towards school- remains financed by DPs, and authority to budget managers at based management, with schools schools have traditionally had little the MoEYS to facilitate expenditure and higher education institutions flexibility in using funding execution, while demanding to enjoying greater autonomy while assign personnel costs to specific being held legally accountable programs and include DP-funded through monitoring and quality projects in the BSPs assurance mechanisms Conduct an assessment of school needs based on size, location, and education level to adjust the distributional formula in the School Improvement Fund Education expenditures on prep- Prepare awareness and commu- Expand the coverage and fund- rimary, primary and secondary is nication campaigns targeting both ing of the current student grant becoming more pro-poor, but is parents and children to further programs to reduce the opportunity still pro-rich, and this is driven by increase enrollment in pre-primary cost of being out school, especially enrollment gaps and primary (where a number of for secondary education children are considered “too young to attend”) Conduct local cost-benefit analysis to determine whether building more schools or reducing access cost through transport, housing options or grant options is most optimal in Northeastern provinces with signifi- cantly lower enrollment As of 2015, about 10 percent of Reallocate teachers to nearby Enforce student-teacher ratios teacher MoEYS salaries went to schools, especially for primary and implement norms that prevent redundant teachers in schools with education (43 percent of schools relocations to schools with excess excess staff, while teacher shortag- with shortage are within 5 km. of teachers es in other schools are equivalent to schools with excess). When not 40 percent of the current payroll possible, provide monetary incen- tives to travel / do additional shift Learning outcomes in Cambodia Increase parent engagement in Introduce the requirement of remain poor across levels. Teacher school-related meetings and en- passing quality enhancing teaching and student behavior are found to hance parents’ awareness on the certificates based on knowledge of impact test scores importance of school attendance the curriculum and classroom be- and homework habits havior “best practices”. Use mobile technologies to better control for teacher absenteeism Source: World Bank staff elaboration. 92 Cambodia: A Public Expenditure Review 5 SUPPORTING AGRICULTURAL DIVERSIFICATION AND RESILIENCE Introduction and motivation of 0.8 percent over the subsequent three years, in D a context of plummeting prices (Figure 49). Adverse weather conditions have also played a role, with a uring 2004-2012, strong agricultural large proportion of paddy and cassava-growing areas growth played a critical role in Cambodia’s experiencing production losses during the 2015/16 El impressive poverty reduction. Especially Niño drought. Due to its large share of harvested area between 2008 and 2012, high international commodity (76 percent) and agricultural value added (approximately prices incentivized farmers and others to expand one-third), the slowdown in paddy production has had their planted area, increase their use of improved the largest proportional impact on aggregate agricultural technologies, and diversify production, all leading to GDP growth. Paddy production grew by an average of increased productivity. This contributed to a remarkable 7.6 percent between 2010 and 2012 then was more decline in the poverty headcount, from 47.8 percent or less flat in 2014 and 2015. Cassava, an important of the population in 2007 to 13.5 percent in 2014. An secondary food and feed crop, saw rapid growth up open trade policy, including allowing the exports of un- through 2012 followed by sharp declines and rubber milled paddy and other raw materials, enabled the quick production, while expanding, has seen dramatically pass-through of international prices to farmers (and falling prices since 2011. Livestock production has been farm laborers). While the process of economic structural affected by disease outbreaks (for poultry) and structural transformation typically involves a declining role for changes (i.e. a shift from draught animal power to land agriculture, the sector still accounts for 26.6 percent of preparation machines). Forestry output has declined GDP and 41.5 percent of employment in Cambodia.65 following the enforcement of logging restrictions and During this period, a Public Expenditure Review exercise a curtailment of new economic land concessions, was undertaken. See Box 7 for a summary of the 2011 measures needed to ensure sustainability. Hence, the Review findings and subsequent developments. agricultural sector as a whole has recently experienced some very strong headwinds, in relation to factors that Since 2013, agricultural growth has slowed previously had spurred growth. significantly in a context of low international commodity prices and adverse weather events. Increased agricultural public expenditures in recent After averaging 4.7 percent during 2006-2012, years do not yet seem to have translated into a Cambodia’s agricultural GDP growth fell to an average 65  Percent GDP from NIS national accounts 2015, percent of employment from CSES 2015. Cambodia: A Public Expenditure Review 93 more resilient and profitable sector, pointing to the going forward. Section 5.2 provides an overview of need for a new strategic approach. Certainly, there public spending trends in agriculture and irrigation, are examples of public programs and DP-supported putting their magnitude into an international context. projects that are having a positive impact on farm Section 5.3 then reviews the functional composition of productivity and incomes. However, these impacts have agriculture spending and the provision of core public tended to be localized, with limited spillovers beyond goods. Section 5.4 focuses on public spending in the areas targeted by the projects and limited impact on irrigation and some strategic considerations to help broader value chains and overall sector performance. A foster higher returns from such investment. Section 5.5 larger return on such programs and projects is needed summarizes the experience with program budgeting as Cambodian agriculture cannot sustainably expand its thus far in MAFF and the Ministry of Water Resource land area or further intensify its fishing efforts, nor can Management and Meteorology (MOWRAM). Finally, it rely upon a return to exceptionally high commodity Section 6 suggests policy options to support agricultural prices. As recognized in the Agriculture Sector Strategic diversification and resilience. Development Plan (ASDP) 2014-18, the sector needs to pivot toward a strategy of sustainable intensification, A few caveats to the findings should be noted. diversification, and value addition, both by strengthening Cambodia’s situation is unusual in that a majority of public the provision of core public goods and by improving agricultural spending is financed by DP organizations, the enabling environment for private investment and with a significant proportion of projects occurring outside services. In this pivoting process, the quality of public of government budgets and/or institutions. In relation agricultural spending and the synergies achieved across to government spending, since program budgeting has programs will become increasingly important. been introduced just recently, there is still little information on program and activity outcomes, which precludes an Drawing from the analysis of public expenditure assessment of agricultural expenditure effectiveness and trends and policies in Cambodia as well as from efficiency. To remedy this going forward, collecting detailed international experience, this chapter identifies information on agricultural program outcomes as defined interventions aimed at boosting agricultural growth in sectoral and budget strategic plans will be essential. Figure 49. Once spurred by rising commodity prices, agricultural growth has fallen as prices have declined 150 6% Commodity price index (2010=100) 130 5% 2010; 100 Agriculture GDP growth 110 4% 90 3% 70 2% 50 1% 30 0% 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 As GDP growth (constant 2000 prices, % yearly) Rice Thai 5%, real index (2010=100) Rubber, SGP/MYS, real index (2010=100) Cassava real price index (2010=100) Source: Rice and Rubber Index data from World Bank Global Economic Monitor (Commodities); Cassava price data from Thai Tapioca Starch Association; Cambodian growth data 2006-2015 from NIS, 2016 MEF estimate. 94 Cambodia: A Public Expenditure Review Box 7 Agriculture and irrigation in the 2011 Public Expenditure Review This box summarizes key findings and recommendations related to agriculture and irrigation in the previous Public Expenditure Review for Cambodia (World Bank, 2011), as well as subsequent developments that are discussed in this chapter. The 2011 PER recommended raising public spending in agriculture, which was significantly lower than in comparator countries (fully implemented). Due to limited fiscal space, the PER recommended increasing government spending only in line with responsible government expenditure, while focusing on the efficiency of spending. Following the report, the RGC increased agriculture (including irrigation) spending by 82 percent from 2010 to 2015, reaching levels similar to those of comparator countries. Meanwhile, overall government spending was kept in line with domestic revenue expansion. The PER recommended that spending increases be directed toward extension, irrigation (particularly secondary and tertiary canals and O&M), research, and rural roads (partially implemented). The PER also noted low wage allocation, at 33 percent of MAFF budget and 17 percent of MOWRAM budget. Subsequently, RGC spending increases have come primarily in irrigation investment and wage increases, with the recent introduction and increase of irrigation O&M budget. Research spending has remained very low but received increased attention in the 2017 budget. While RGC extension spending has remained low, DPs have allocated large funding shares to this activity. DPs covered 57 percent of agriculture spending in 2010, with this share increasing to 66 percent in 2015. During 2013-2015, 73 percent of DP agriculture funding went to irrigation construction/rehabilitation. While spending has increased in most of the recommended areas, MOWRAM still suffers from limited water management funding and capacity as well as insufficient staffing. The PER also identified needed institutional changes in extension and research to clarify responsibilities and improve institutional functionality (partially implemented). The RGC began such a process for research in 2017. The extension policy was recently introduced, outlining institutional requirements for an effective extension service. The RGC will now focus on successful implementation of the policy. In terms of policy and planning, the PER noted that the lack of coordination between policy, programs, and the budget process needed to be addressed (partially implemented). The process of strengthening the program budgeting process is ongoing, and many improvements have been made. At this stage, the Ministries face different challenges. For MAFF, challenges remain in DP funding inclusion, sub-program consolidation, outcome indicator strengthening and collection, and lack of clear prioritization. MOWRAM’s immediate challenges are lack of realism in BSP preparation, limited alignment between programs and the water law that sets out its mandate, unclear sub-program titles and objectives, limited expenditure inclusion and breakdown in strategic documents, unrealistic and unclear outcome indicators, and discrepancies between MOWRAM closing reports and MEF Treasury reports. For implementation efficiency, the PER 2011 produced benefit-cost ratios to indicate that irrigation investment had fallen short of potential, and it recommended refocusing irrigation spending (partially implemented). The PER pointed to problems such as incomplete rehabilitation, missing O&M funding, and insufficient attention to secondary- and tertiary- level systems. While allocations to lower-level irrigation systems and O&M have increased, functionality concerns remain with regard to irrigation schemes, in terms of both O&M and design. Cambodia: A Public Expenditure Review 95 Box 7 The PER 2011 recommendations also included the need for a monitoring system to allow for assessment of extension efficiency as well as for other programs (not implemented). The PER demonstrated the potential return to extension services, but the effectiveness of current extension arrangements is uncertain. Insufficient intermediate outcome data collection in both ministries still prevents specific program assessments. As such, one cannot determine how public spending and other agriculture inputs translate into agriculture sector outcomes. Finally, the PER 2011 cited climate, substantial available land area, and surface water as key strengths in the Cambodian agricultural sector, but new sources of agricultural growth are needed going forward. Around 60.5 percent of the subsequent growth (2010-2012, plus 229 percent of 2014 growth) came from land expansion, with this opportunity now largely exhausted. In the current situation, intensification and diversification have therefore become the potential sources of growth, making the need for quality infrastructure and support services even more pressing. Public spending trends in spending has nearly doubled, while DP contributions agriculture and irrigation have remained largely stable since 2010. DP funding for non-irrigated agriculture now seems to be on a downward trend. Nonetheless, the total (MAFF plus Public expenditure in agriculture has increased DP) agricultural budget in 2015 was around KHR 365 significantly over the past decade. Looking first at billion, a figure nearly double that of the late 2000s—a the MAFF, its budget increased more than threefold in period in which there was some consensus that there nominal terms and twofold in real terms between 2007 was underinvestment in Cambodian agriculture (World and 2016 (Figure 50, left panel). As a share of GDP, the Bank, 2011). MAFF budget grew from 0.18 percent to 0.26 percent over the same period, while its share in the total budget When irrigation spending is included, total agricultural remained at about 1 percent. The allocations to MAFF outlays triple, mainly thanks to the scaling up of DP were boosted significantly in the 2017 budget, reaching funding for irrigation schemes in recent years (Figure 1.28 percent of total budget and 0.3 percent of GDP. 51, left panel). Government spending in MOWRAM has traditionally been higher than that of MAFF, although DP-funded agriculture-related spending also more recent increases in the latter have narrowed the gap. than doubled in nominal terms between 2007 and DP spending in irrigation has been about three times the 2015. DP-funded spending in agriculture (excluding amount spent on other aspects of agriculture in recent irrigation) jumped from KHR 90 billion in 2007 to KHR years. When taking irrigation into account, DPs financed 183 billion in 2015 (Figure 50, right panel), remaining more than two-thirds of total public agricultural spending, higher than government funding. Estimates for 2016 and irrigation-related spending also represents around indicate that the government may have surpassed two-thirds of total public spending in the sector (Figure DPs however. It should also be noted that the two are 51, right panel). not directly comparable.66 In real terms, government 66  Some DP projects are implemented by or through MAFF, yet the majority are implemented directly by DPs or their subcontractors. In some cases, considerable resources have gone to cover the costs of expatriate managerial and technical staff, including the overhead costs of their home institutions. Grants to non-state actors are an important element of some DP-supported projects. Hence, some DP spending should not be regard- ed as ‘equivalent’ to that of the RGC. Deciphering whether DP spending is/is not of a similar nature to MAFF is not possible given the absence of systematic information on the economic composition of DP (non-capital) spending. For illustrative purposes, therefore, DP agricultural spending was simply combined with that of the MAFF budget to obtain the “total agricultural budget.” 96 Cambodia: A Public Expenditure Review Figure 50. Government spending in agriculture has more than doubled in real terms since 2007, while DP spending in agriculture (excluding irrigation) has remained relatively stable in real terms since 2010. Nominal and real expenditure trends in agriculture spending Government spending Development Partner Spending 300,000 300,000 250,000 250,000 MAFF budget, million Riels DP budget, million Riels 200,000 200,000 150,000 150,000 100,000 100,000 50,000 50,000 0 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2007 2008 2009 2010 2011 2012 2013 2014 20152016e Nominal Real Nominal Real Note: e.2016 estimated DP expenditure. Source: MEF of Cambodia, CDC website for development budget. Figure 51. DP spending in irrigation has increased dramatically (left side), and irrigation now makes up two-thirds of total agriculture spending (right side) 1,000,000 1,400,000 MOWRAM budget, million Riels 1,200,000 800,000 Budget, million Riels 1,000,000 600,000 800,000 400,000 600,000 400,000 200,000 200,000 0 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016e 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016e Government Donors Irrigation Other agriculture Note: e. 2016 estimated DP expenditure. Source: MEF of Cambodia, CDC website for development budget. Although investments in irrigation can potentially maintenance (O&M) and tertiary canals, limiting the play a significant role in supporting sustainable connection to the fields. Investments have also been agricultural intensification and diversification completed without being fully operational, lacking in Cambodia, it is too soon to determine the a reliable year-round water supply and functioning effectiveness of the sharply increased irrigation operational arrangements, thereby failing to allow spending in recent years. Much of this investment for year-round agriculture. Potentially further limiting has been made at high (primary and secondary) levels, investment returns, coordination between MOWRAM with insufficient attention given to operations and (for water delivery) and MAFF (for agricultural service Cambodia: A Public Expenditure Review 97 delivery) has not been strong. Finally, the focus of recent be noted that for a few other comparator countries—i.e. investments appears to be almost entirely on rice, as Vietnam and the Philippines—DP spending in agriculture opposed to higher-value crops which could deliver is likely to be substantial, but not included in this table. a higher return on the infrastructure investment and support an effective response to changing domestic In summary, agricultural spending has increased dietary patterns in Cambodia and neighboring countries. significantly in recent years but remains dominated Given their scale and importance, irrigation investments, by DPs. From the country’s fiscal resource allocation their functionality, and their linkages with the evolving perspective, the agricultural sector is already a strategic orientations for the agricultural sector should significant recipient of public funds, especially when be monitored closely. irrigation is taken into account. The continued high reliance on DPs to finance agricultural programs Overall, government spending in agriculture and calls for a more proactive and strategic engagement irrigation, at 0.5 percent of GDP in 2014, is similar to among public actors and DPs during project selection, that of other middle-income economies. As a share design, and preparation to ensure a strong alignment of agricultural GDP (1.9 percent), it remains below the with government programs and priorities, as well as levels observed in other countries (Table 28), mainly due coordination with other ongoing activities. The next to the fact that agriculture still represents a large share two sections delve in more detail into the economic of the economy in Cambodia while the comparator and functional allocation of public resources going into countries have already experienced deeper economic irrigation and agriculture. See Box 8 for an overview structural changes. When DP spending is added, total of international experiences regarding agricultural spending triples, reaching 1.8 percent of GDP. It should spending to frame the following discussion. Table 28. Government spending in agriculture as a share of GDP is similar to comparator countries, although agricultural budget as a share of agricultural GDP is relatively low Agricultural budget as a share of total and agricultural GDP, 2014 Agricultural Agricultural budget Agricultural budget budget as a share as a share of agri- as share of Govern- of total GDP, % cultural GDP, % ment revenue, % Chile 0.30 8.49 1.4 Turkey 1.25 15.00 3.4 South Africa 0.23 9.05 0.8 Vietnam 0.24 1.28 1.1 Brazil 0.48 5.89 1.9 Philippines 0.48 3.73 3.2 China 0.66 7.48 4.2 Cambodia (MAFF) 0.26 0.89 1.6 Cambodia (MAFF and MOWRAM) 0.54 1.88 3.3 Cambodia incl. DP (Agriculture) 0.59 2.04 Cambodia incl. DP (Agriculture & Irrigation) 1.83 6.33 Source: Agriculture spending: MEF for Cambodia, OECD for comparator countries. Government revenue: World Development Indicators. 98 Cambodia: A Public Expenditure Review Box 8 Public Expenditures and Agricultural Performance: Lessons from International Experience Over the years, the World Bank, International Food Policy Research Institute, and other actors have examined the linkages between various forms of public spending and the performance of the agricultural sector. Several broad lessons can be draw from these analyses, all of which are pertinent for Cambodia. Lesson #1: Public expenditures matter for agricultural growth. A study of ten Latin American countries found that a 10 percent increase in agricultural public expenditures lifted agricultural growth by 0.6 percent during 1985-2000 (Lopez, 2005). The rationale for public investments is derived from two fundamental sources: (i) economic inefficiencies resulting from market failures and (ii) inequalities in the distribution of goods and services. The benefits from public goods for agriculture, which stimulate growth, can be grouped into four pathways (World Bank, 2016): a. Generating technology: Investments in research and development (R&D) are among the most important public goods and a critical component of agricultural public spending. The returns to R&D include benefits not only to the farm sector but also to the food industry and consumers. The private sector tends to underinvest in agricultural R&D, requiring governments to correct this market failure although the actual conduct of the research need not be by government agencies. b. Disseminating knowledge and building more human capital: Human capital-enhancing effects can be associated with public spending on extension, training, and information services that transfer knowledge and skills to those engaged in farming. c. Reducing transaction costs: Rural roads are a critical element of public infrastructure for agricultural growth. Similarly, institutional investments to overcome barriers to collective action and reduce transaction costs to improve collection, storage, input and output quality control, and price information can optimize supply chain management. d. Attracting private capital: The crowding-in effects of agricultural public spending on private capital occur when public and private investments are complements in production. An example is public investment in large irrigation infrastructure such as dams and canals, which then make it profitable for farmers to make small on-farm investments in water management and a wider range of production technologies. Lesson #2: Not all public expenditures are productive. Many empirical studies find that aggregate spending has a weak impact on agricultural growth (IFPRI, 2012), implying that all investments are not equal. Governments sometimes spend on things that are not public goods. In addition, even when clear failures exist in particular markets, government spending will not necessarily improve the situation. Strong empirical evidence comes from Latin America, where a study found that agricultural public spending on public goods was much more productive than public spending on private goods (Lopez, 2005). The study found that reallocation of 10 percentage points of total public expenditures from subsidies to public goods increased per capita agricultural income by 2.3 percent. This was obtained without increasing total expenditures. Another study found that different types of spending have had varied impacts on agricultural growth and rural poverty reduction in India, with returns also changing over time (Fan, Gulati, & Thorat, 2008). The study highlighted the continuously high return to public goods (particularly R&D) compared to other measures, something later echoed by Rashid et al. (2013),who found that technology adoption in India, Bangladesh, Pakistan, and Indonesia was far more influenced by effective agricultural R&D, road improvements, and improved irrigation services than the provision of fertilizer subsidies. Cambodia: A Public Expenditure Review 99 Box 8 Lesson #3: The economic composition of spending must be well-balanced to achieve high impact. For example, when an entire budget for extension services is spent on salaries and there is no fuel for motorbikes and vehicles for farm visits, high spending on extension is not effective. Likewise, inadequate spending on O&M for rural roads and irrigation investments reduces the economic value of these assets, leading to higher budget outlays in the longer run. The lesson from around the world is that the agricultural budget needs to be well-balanced across subcategories (wages, nonwage recurrent, and capital expenditures) to make agricultural programs effective (World Bank, 2011). Lesson #4: Quality of implementation affects the outcomes of even justifiable public programs. When delivering public programs, attention must be given to: unit costs and value for money, implementers’ capacity, a targeting strategy to identify and reach intended beneficiaries, collaboration of various departments within and between the ministries, division of labor between central and local governments, and establishment and use of effective monitoring and evaluation. Monitoring the absorption capacity of institutions through budget execution is also important. In the case of Cambodia, the effectiveness of, and synergies between, DP-supported investments and the relationship of these to national and provincial programs are particularly important. Lesson #5: Investments in public goods combined with better policies and institutions bring about the best results. Improvements in the policy environment through trade and regulatory reforms augment public spending by enhancing incentives for producers and innovators to take advantage of public goods, thereby crowding in private investments. In contrast, distortions such as input, credit, or output subsidies usually crowd out private investments, while restrictions on marketing, trade, land use, and other functions often dull the incentives to invest. Agricultural spending allocations flat in real terms until 2017, when allocations were more and the provision of core public than doubled in real terms with significant investments goods in laboratory facilities and equipment as well as some investment in food storage facilities. In terms of economic composition, Purchases and services accounted for 80 percent government spending in agriculture of total non-wage recurrent budget allocations in seems to be well-balanced overall 2017. Within purchases (Chapter 60), allocations to supplies and food and agricultural goods increased Government spending in agriculture has increased in a contained manner, at yearly rates of 3.4 and 5.9 significantly in recent years, both in terms of wages percent, respectively, during 2010-17 (Table 29). Under and other recurrent spending. Similar to other services (Chapter 61), travel allowances and vehicle rent ministries, between 2013 and 2017, the public payroll (growing around 20 percent a year) and communications at MAFF grew at an average rate of 8.2 percent in real (15 percent) experienced very significant increases. terms. This has been driven by rising public sector Training fees also surged, and overall, this suggests that wages rather than significant increases in personnel frontline service providers are receiving better training numbers. Over the same period, non-wage recurrent and more operational funds to perform their day-to- spending rose at an average rate of 8.4 percent in real day duties which, accompanied by rising wages, is terms. Government-financed capital spending remained 100 Cambodia: A Public Expenditure Review Figure 52. Wages and other recurrent spending in agriculture have increased significantly in recent years Economic composition of MAFF budgets (mill constant riel), 2010-2017 180,000 160,000 140,000 Million Real Riels 120,000 100,000 80,000 60,000 40,000 20,000 0 2010 2011 2012 2013 2014 2015 2016 2017 Wages Non-wage recurrent Capital Note: Adjusted budget for 2010-2015, approved budget for 2016-2017. Source: MEF Treasury Accounts. expected to result in improved outcomes. However, the Public spending is geared toward funding maintenance and repair budget and research activities public goods, although Cambodia have seen more modest increases in recent years (at may be underinvesting in the areas of 2.3 percent and 5.6 percent, respectively). agricultural research and disease control Overall, the economic composition of the MAFF This analysis of the functional composition of public budget and its major programs seems to be well- spending aims to assess whether public funds are balanced between wage and non-wage recurrent allocated to priority functions, consistent with the expenditures.67 In 2010-16, wages represented around country’s and sector’s strategic objectives. Global one-third of total MAFF spending (35.3 percent), with the experience shows that allocation decisions matter, remainder going to non-wage recurrent expenditures because some functions and activities have higher rates (53.5 percent) and capital expenditures (11.2 percent). of return than others (see for example Box 9 on the The shares have been fairly stable, and wage increases topic of fertilizer). These ‘returns’ can relate to aggregate do not seem to have crowded out recurrent spending sectoral growth, growth among focal commodities or (Figure 52). The budget allocations for training, sub-sectors, impacts on poverty, contribution to health travel, and communication have increased over time, and nutritional outcomes, support for lagging regions, facilitating the operation of frontline service providers. or other goals. Unfortunately, during preparation of this Spending in O&M may need to increase more rapidly PER, data that would enable an assessment of linkages going forward, in line with the boost of capital spending between agricultural spending in specific activities allocations to MAFF included in the 2017 budget. Thus and outcomes was not available. Thus, this section far, most capital-intensive agriculture-related programs focuses on analyzing public expenditure allocation have been implemented by other ministries, such as across functions during the first two years of program MOWRAM and Ministry of Rural Development (MRD). budgeting implementation (2015 and 2016). 67  A caveat is that insufficient information on DP funding allocations to agriculture by economic classification limits the ability to draw conclusions. Cambodia: A Public Expenditure Review 101 Table 29. Budget allocated to Chapters 60 and 61 and their sub-accounts, mill Riels (nominal) 2010 2011 2012 2013 2014 2015 2016 2017 Percent Avg. of total yearly real Non-wage change, recurrent, percent 2017 Chapter 60: Purchases 18,508 22,862 24,021 28,611 27,572 25,651 28,654 32,102 27.0 5.2 Maintenance and sup- 5,168 5,014 5,999 7,118 7,629 7,001 7,656 7,899 6.6 3.4 plies Food and agricultural 5,111 4,403 4,932 6,712 6,865 6,682 8,107 8,917 7.5 5.9 products Office supplies 2,559 2,688 2,445 3,113 3,368 4,330 4,113 4,127 3.5 4.5 Furniture, equipment 2,819 7,200 6,761 7,750 6,278 4,119 4,522 5,677 4.8 16.5 Utilities 2,597 3,245 3,557 3,588 3,150 3,236 3,486 3,833 3.2 2.9 Uniform and 254 311 327 332 283 283 761 1,618 1.4 37.9 decorations Chapter 61: Services 24,732 24,880 28,362 28,761 34,609 46,163 54,764 63,964 53.8 11.6 Maintenance and re- 10,671 11,215 10,600 12,252 12,266 12,709 13,199 15,376 12.9 2.3 pairing Local allowance/travel 5,459 5,573 6,791 7,180 9,156 15,979 21,571 23,958 20.1 21.7 Communication & 3,845 4,163 5,687 4,239 6,088 8,860 9,548 10,943 9.2 15.2 dissemination Vehicle rent 790 345 799 991 1,813 2,132 1,438 1,307 1.1 19.3 International allowance/ 930 1,000 1,165 1,200 1,528 1,893 2,154 3,459 2.9 18.1 travel External staff 787 564 583 1,042 1,751 1,827 1,673 1,515 1.3 12.0 Research and studies 1,107 1,246 2,022 997 990 1,080 1,077 1,394 1.2 5.6 Training fees 365 128 115 204 288 957 3,362 4,804 4.0 76.2 Other contracted 778 644 601 657 730 722 720 1,205 1.0 5.6 services (rent, transpor- tation fees, media, etc.) TOTAL Chapters 43,240 47,743 52,384 57,373 62,182 71,815 83,418 96,066 80.8 8.7 60 and 61 TOTAL Non-Wage 58,808 64,626 69,534 75,767 79,911 90,133 103,284 118,969 100.0 7.2 Recurrent Budget Source: MEF Treasury Accounts. 102 Cambodia: A Public Expenditure Review Table 30. While allocations in some areas have fallen short of planned ASDP spending, the government agricultural budget seems well-aligned overall Total recurrent agricultural budget by program, ASDP versus budgeted, mill Riels Programs 2015 2016 ASDP Budget DPs ASDP Budget DPs Program 1: Improved agric. produc- 46,184 30,531 112,682 46,877 36,487 87,784 tivity, diversification, and commercial- ization Program 2: Support to animal pro- 29,743 14,491 10,879 32,717 16,712 14,042 duction and health Program 3: Sustainable fisheries 23,569 19,979 15,405 25,457 22,584 22,723 resource management Program 4: Sustainable forest re- 34,061 23,180 5,097 45,894 26,287 1,431 sources and wildlife management Program 5: Support to institutions, 32,716 62,116 39,182 35,624 85,954 53,642 support services, and human re- sources TOTAL 166,275 150,297 183,246 223,886 188,025 179,622 Source: ASDP, recurrent expenditure requirements, World Bank staff estimates based on the data from MEF and CDC. Allocations by program have been falling short of DP funding is concentrated in programs related those reflected in the ASDP but are well-aligned to crop production and institutional support. DP overall. Government budget allocations to Programs 1 funding in crops (and agricultural extension) is more to 4 seem to have fallen short of ASDP planned spending concentrated than that of government, with Programs 1 (Table 30). The resource gap is especially significant and 5 receiving 83 and 79 percent of total DP spending for Program 2 (animal production) and Program 4 in 2015 and 2016, respectively. The reduction in DP (forestry), where additional DP funding has been limited funding under Program 1 from 2015 to 2016 largely and, in some cases, insufficient to compensate for low reflects the end of the USAID HARVEST project in 2015, government spending. At the same time, DPs alone with HARVEST II making its first disbursement in 2017. have been funding three times the ASDP planned As such, it is likely that 2017 will see a rebound in this outlays in Program 1 (agriculture). Budget allocations to area. The limited DP support to animal production and Program 5 (institutional support and human resources) forestry may be due to a combination of concerns about have been significantly higher than those estimated in governance (i.e. illegal logging) and implementation the ASDP, which may simply reflect that the payroll and risks. Otherwise, these areas—and especially those training of MAFF staff have been lumped in Program 5 related to sustainable natural resources management— (regardless of whether they are working on agriculture, are entry points for significant development support fisheries, or administration). Thus, the gaps in budget programs elsewhere in Southeast Asia. allocations vis-à-vis ASDP for the other programs would be smaller, and the government budget is well-aligned Among cross-cutting functions, transfers to overall in terms of distribution across programs. provincial programs, planning and management, Cambodia: A Public Expenditure Review 103 and human resource development account for research, inspection services and laboratories, two-thirds of the total, while allocations for and extension represented 18 percent of recurrent activities promoting agricultural diversification budget in 2016. Considering the ASDP objective have been low. Table 31 shows the result of grouping of promoting agricultural diversification, funding to sub-programs and activities that respond to the same support productivity, post-harvest management, or functions even if they are in different programs (e.g., other advances for commodities other than rice and education in agriculture and in livestock). Nearly 30 rubber has been small and very likely inadequate. For percent of the MAFF recurrent budget has gone example, funding to support animal (2.4 percent) and toward provincial programs, consistent with the horticultural (0.8 percent) products remains limited aim of increasing the decentralization of essential despite growing demand from consumers and tourists services. 68 Another quarter has been devoted to and despite the need to boost the quality of domestic sector planning and management capacity. Functions production and manage biosecurity risks. Growing associated with the provision of core public goods demand for higher-value foods has mainly been met (highlighted in orange in the table) such as education, by imports from Vietnam, Thailand, and elsewhere. Table 31. Functional composition of the MAFF recurrent budget Functional category MAFF 2015 2016 2015 2016 (mill riel) (mill riel) (% of MAFF (% of MAFF recurrent budget) recurrent budget) Provincial programs 44,033 56,965 29.3% 30.3% Planning and management capacity 37,322 43,667 24.8% 23.2% Human resources 18,952 28,615 12.6% 15.2% Education 8,649 11,431 5.8% 6.1% Production support to Fisheries 1,627 2,226 1.1% 1.2% Production support to other products 6,840 7,113 4.6% 3.8% Research 6,464 7,632 4.3% 4.1% Inspection services 6,464 6,454 4.3% 3.4% Extension and farmer organizations 3,437 4,888 2.3% 2.6% Animal health services 4,108 4,453 2.7% 2.4% Rubber programs 3,969 4,420 2.6% 2.4% SPS and laboratories 2,776 3,197 1.8% 1.7% Agro-industry development 2,361 2,954 1.6% 1.6% Rice 2,095 2,599 1.4% 1.4% Horticulture 1,200 1,412 0.8% 0.8% Total 150,297 188,025 100% 100% Source: MEF and MAFF, based on sub-program level data. Note: Cross-cutting public goods highlighted in red. 68  In some cases, DPs have provided very substantial supplements to the provincial budgets to implement dedicated programs. In particular, the IFAD-supported ASPIRE project cooperates closely with Provincial Departments of Agriculture Forestry and Fisheries (PDAs) and aims to develop an enhanced model of agricultural extension services, both with regard to PDA staff capacity and the content of services. Another DP project that works closely with PDAs is the CAVAC project funded by Australia. 104 Cambodia: A Public Expenditure Review Table 32. DP Expenditure in Agriculture by function, mill riels 2013-2015 Categorization 2013 2014 2015 2016 share share 2015 2016 Food Crops, Food Security, Nutrition 28,694 38,227 35,211 40,375 19.2% 22.5% Agricultural Sector Policy and Man- 59,086 18,574 20,201 23,818 11.0% 13.3% agement Extension Services 35,339 87,888 56,013 23,544 30.6% 13.1% Fisheries 9,102 7,680 15,405 22,723 8.4% 12.7% Agricultural Finance Services 2,790 17,102 17,409 16,332 9.5% 9.1% Livestock and Veterinary 8,078 4,299 10,879 14,042 5.9% 7.8% Education, Training 17,905 12,586 10,553 8,973 5.8% 5.0% Industrial and Export Crops 3,556 5,291 2,245 4,153 1.2% 2.3% Agricultural Inputs 1,598 - 573 2,342 0.3% 1.3% Forestry 5,542 13,689 5,097 1,431 2.8% 0.8% Agro-industry - - 695 874 0.4% 0.5% Post-harvest - 367 536 162 0.3% 0.1% Other 3,329 13,963 8,428 20,851 4.6% 11.6% Total 175,018 219,664 183,246 179,622 100.0% 100.0% Source: CDC Database. In terms of functional areas, DP spending has pertinent to the goal of agricultural diversification— been focused on crops, food security, and has been very limited. nutrition as well as the provision of some public goods such as extension and finance services Government spending on extension services has (Table 32). These foci, together with agricultural been low, however this is consistent with the aim of sector policy and management, accounted for nearly supporting a pluralistic system of advisory services 70 percent of total DP spending in 2015, and an and has been supplemented by heavy DP funding. estimated 58 percent in 2016. Historically, most The recent National Agricultural Extension Policy DPs have been supportive of extension and food (2015d) called for a pluralistic, demand-driven system security and nutrition initiatives, although attention in which government provides an enabling environment is increasingly shifting toward funding commercial for a broad range of (private, NGO, academic) service competitiveness and value chain development. providers, in line with international best practices (Box In relation to this, agricultural finance services 10). In this context, government spending on extension have experienced the most dramatic increase in increased by 50 percent between 2015 and 2016, DP funding. Meanwhile, funding to support the driven by wages and service spending, but remains development of industrial and export crops or to limited, leaving space for other actors. DPs financed 96 support improved post-harvest management—both percent of extension spending in 2015. Cambodia: A Public Expenditure Review 105 Box 9 Promise and perils of fertilizer subsidies (excerpt from Annex 2 of the PER agriculture discussion paper) One area of potential relevance to Cambodia is input subsidies. The benefits of judicious use of chemical fertilizers are well-known and played an important contributing role to the productivity and food security gains during the Green Revolution in Asia. Yet in many parts of the world less-than- optimal rates of fertilizer use occur, and the mix of nutrients applied often does not meet specific local needs based on the farmer’s soil and other growing conditions as well as the crops being grown. Various factors contributing to this situation include: • Risk: Agriculture is a risky endeavour, given the impacts of weather and uncertainties in both output and prices. Farmers may reduce their financial risks by using levels of inputs lower than the level that could increase their profitability. • Knowledge: Farmers may not be aware of the benefits of fertilizer use or may lack the information or skills to apply it correctly. • Non-affordability: Extremely poor farmers may lack the money to purchase fertilizer, especially at the time when its application is needed. • Non-accessibility: Due to high transport costs or other logistical barriers and perhaps the very limited local market potential for distributors, some farmers may not have ready access to fertilizer. • Underdeveloped supply: Limited domestic demand may contribute to low investment in fertilizer manufacturing and distribution systems, leading to overall low levels of fertilizer available for farmers anywhere in the country. Where such conditions apply, fertilizer subsidies can potentially encourage technology adoption and diffusion by reducing the initial (financial) risks and the costs of learning a new technology (World Bank, 2008). They can potentially contribute to a variety of objectives (Timmer & McCullough, 2010), including increasing overall agricultural productivity, enabling pro-poor growth, and providing a safety net for the extreme poor. Sometimes, subsidies are justified simply on the basis of wanting to achieve higher output of one or many crops. Fertilizer subsidies are also a very visible way for a government to demonstrate its support for agriculture even where the economic and social benefits are modest or unknown (Jayne & Rashid, 2013). Where fertilizer subsidies are introduced, World Bank (2008) advocates that the programs apply a set of ‘smart subsidy’ principles. That is, programs should be: • Focused, targeting farmers who were not utilizing fertilizer, yet would find it profitable to do so; • Market supporting, using the production/distribution systems of the private sector to deliver the subsidy; • Synergistic, implementing the subsidies as part of multi-dimensional efforts to improve technology use and agricultural market development; and • Temporary, having a clear ‘exit’ strategy based upon monitored results. In practice, several of these principles have been difficult to apply. In addition to the above, fertilizer subsidies need to be fiscally responsible and not crowd out investments in important public goods— such as agricultural research, animal disease control, and others—which have been demonstrated in many countries to have more significant and sustainable impacts on agricultural productivity. Ideally, the design of a fertilizer subsidy should reflect a deep analysis of the underlying constraint, information gap, or market failure that is resulting in the less-than-optimal levels of fertilizer use. 106 Cambodia: A Public Expenditure Review Box 9 This is not always done, and where agriculture features considerable heterogeneity in growing conditions and farm types, applying a single technical solution or subsidy delivery mechanism can be too blunt. In Sub-Saharan Africa, responses to fertilizer subsidy schemes have been found to be generally disappointing but with much higher gains in irrigated areas than in rain-fed areas (Jayne, Mather, Mason, & Ricker-Gilbert, 2013). A conclusion of many studies has been that poor and declining soil quality has substantially reduced the efficiency with which food crops utilize fertilizer nutrients. In Asia, following the Green Revolution, studies are finding adverse relationships between farmer yields and very high chemical fertilizer usage. In China, past subsidies have helped create a robust fertilizer industry and have contributed to enormous increases in fertilizer use and progress in raising food output. However, the large subsidies, combined with deficiencies in education and extension efforts, have also resulted in overuse and massive environmental problems including air and water pollution (Li, et al., 2013). In India, unbalanced fertilizer use, stemming from high subsidy rates for nitrogen (in the form of urea), has been associated with negative yield patterns (World Bank, 2014). A common problem in Asia is untested soils, leaving a high likelihood that many farmers are applying suboptimal or even damaging mixes of nutrients. In these cases, a fertilizer subsidy does not contribute to more informed decision-making by farmers and in some circumstances, may even be incompatible with strategies to promote organic or other forms of ‘green’ agriculture. These findings may be instructive for Cambodia, where the bulk of cultivation is, and will continue to be, rain-fed and where rising land pressures, more continuous cultivation, and soil degradation are emerging problems. While increased use of synthetic fertilizers could be beneficial in some areas, the challenge is not simply to increase. Equally important would be efforts to better educate and advise farmers about agronomic practices to rebuild soil organic matter and to practice inter-cropping or crop rotation to help fix nitrogen and otherwise help restore the responsiveness of soil to fertilizer applications. In lieu of fertilizer subsidies, a comprehensive soil fertility management program could be implemented involving soil testing services, more specific fertilizer blends appropriate to farmers’ specific conditions, investment in drainage, and restoration of soil organic matter through various conservation farming practices (e.g., minimum tillage, use of green manures). Tying this to better water management would have an even bigger impact. In other areas, however, public funding may not be the International Rice Research Institute, MAFF, and sufficient to support a critical mass of activity, as the Cambodia Agricultural Research and Development in the case of agricultural research. At 0.2 percent Institute. Other research support comes from the long- of agricultural GDP, public spending in research in standing cooperation with the Australian Centre for Cambodia is well below the world average (0.9 percent). International Agricultural Research,69 as well as support Similarly, the number of researchers per 100,000 on a smaller scale from the IDRC-Canadian International farmers is low at 5.6, compared with neighboring Food Security Research Fund and the U.S.-funded countries such as Vietnam (12.5) and Lao PDR (9.3) Feed the Future Innovation Labs. Overall, DP funding as well as countries at comparable levels of GDP per for research is estimated to be twice that of government capita (Figure 53). DP spending on agricultural research funding, yet the total would still remain well below has been higher than that of government. There are 14 international standards. active programs involving collaborative work between 69  The ACIAR cooperation project has run from 2007, with average yearly spending of 2,659,399 USD, including total counterpart funding of 1,073,663 USD from the RGC between 2007 and 2015, according to the CDC database. Cambodia: A Public Expenditure Review 107 Box 10 Reforming the goals, delivery, and financing of agricultural extension Agricultural extension has evolved significantly as a concept and practice since the 1970s when it spread widely to developing countries. At the time, agricultural extension was by and large a public- sector service focused on raising yields and productivity by providing technical training to farmers. Today, the vision for effective agricultural extension is less that of a public technology transfer service than that of a pluralistic system of public and private sector actors offering technical and non-technical services, not only to help raise farm-level productivity but also to improve business, health, social, and environmental outcomes in the sector more broadly. This type of thinking is embedded in Cambodia’s National Agricultural Extension Strategy. Extension-plus, as it is sometimes referred to, is focused as much on the brokerage of relationships and the facilitation of reciprocal knowledge flows as it is on the direct, one-way provision of information and technology from knowledge institutions to end users. In its broader quest to improve sector performance by boosting the capacity to innovate, it is meant to enhance value chain actors’ access to knowledge, information, and technology and their interactions with research, education, agribusiness, and other institutions. In the same vein, enhanced extension systems strive to help value chain actors develop their own technical, organizational, and management skills as well as the capacity to continuously absorb new knowledge and technology. Extension-plus tends to involve multi-directional information flows and involve multiple disciplines, methods, and stakeholders. Diverse circumstances—in terms of the targeted clients, the types of advice being provided, and the structure and capabilities of value chain players—call for the use of various models for the delivery and financing of farm advisory services. For example, government may finance public advisory services in the field, or alternatively, could contract private companies, NGOs, farmer- based organizations, or others to perform this function. Private companies may have an extension need and contract staff from the public advisory service or from an NGO. Otherwise, they can embed advisory services into their input provision and/or product marketing functions. Figure 53. Cambodia’s public agricultural research spending and staff are well below average 200 Researchers (FTEs per 100,000 farmers) 180 160 140 120 100 80 60 40 KHM, 2014 20 0 0 0.5 1.0 1.5 2.0 Research spending (as a share of AgGDP, %) Source: ASTI database. International Food Policy Research Institute (IFPRI). 108 Cambodia: A Public Expenditure Review Finally, government efforts in the areas of plant equivalent to only USD 2.2 million. This is unlikely to be protection, animal disease control, food safety, adequate for developing and maintaining the necessary and other regulatory services may be falling facilities, equipment, and human capital to provide highly short of needs. The adequacy of spending—and professional services in these areas. It is not uncommon the functionality of public systems—in these areas for countries to underinvest in these and other regulatory would need to be reviewed closely, not only taking functions, only to incur much higher direct and indirect current risks into account but also anticipating future costs following an avoidable disease outbreak, food challenges, given demographic and changing dietary safety scare, or SPS-related market access restriction. trends. In 2016, the combined budgetary allocations for Investing in preventative measures and supporting ‘good’ inspection services, sanitary and phytosanitary (SPS) agricultural, manufacturing, and hygiene practices can measures, and the laboratory testing sub-program were help Cambodia avoid high future costs (Box 11). Box 11 Food safety: spending public resources wisely With demographic and dietary changes, the assurance of food safety will become a growing policy concern in Cambodia. Increased investment will be needed to strengthen core capacities in the public sector to provide effective regulatory oversight, even as the food system players themselves are given primary responsibility for ensuring that food safety. International experience points to several principles that can enable governments to prioritize actions and make strategic use of public expenditures in this area. For example: • Proactive prevention. Governments should give precedence to the prevention of hazards over the inspection of finished products. While this approach has been spearheaded in developed countries, it is highly relevant to emerging market countries. The hazard analysis and critical control points (HACCP) approach is one of the best known and most widely adopted prevention approaches. HACCP is a management system in which food safety is addressed through the analysis and control of biological, chemical, and physical hazards from raw material production procurement and handling to manufacturing, distribution, and consumption of the finished product. Its principles require that (identified and realistic) food safety hazards be prevented, eliminated, or reduced to acceptable levels. • Risk profiling. One of the defining pillars of proactive prevention (i.e. HACCP) is the principle of hazard monitoring. One practical way to do this involves auditing and assigning risk profiles to enterprises. Risk profiling generally entails a shift from the retrospective recording of food safety breaches to the more proactive assessment of how likely future problems are. Risk profiling generally places less emphasis on the safety of end-products than on that of business facilities and practices. Ideally, risk profiling also takes into account risks that arise through sourcing and primary production of the foods that firms use. • Influencing behavior. Some of the most effective uses of public resources are ones that seek to influence the behavior of farmers, firms, and consumers in ways that help prevent unsafe food. For example, it is common for public extension services to develop and disseminate agricultural best practices that are designed with food safety in mind. Also common is having public agencies regularly publish protocols and instructions on safe food handling in the manufacturing and service sectors. In some countries, (subsidized) training and capacity building programs are offered to SMEs. Not only do these impart knowledge on safe food handling and the serious risks associated with departures from protocol, but they also (presumably) build a professional identity and ethic that can be instrumental in changing the behavior of individuals employed in firms, even in the absence of public scrutiny. Consumer education (including through schools) is typically highly cost-effective in improving food safety outcomes. Cambodia: A Public Expenditure Review 109 Ensuring returns to investment in basin management, irrigation system management, irrigation and sustainable water use; and to support farmer water user committees (FWUC). This section focuses on irrigation, which makes up the bulk of capital spending The focus on irrigation scheme on agriculture, but it should be recognized that for rehabilitation and O&M spending irrigation system development to be successful, the has increased, but ensuring returns complementary functions and services discussed earlier to investment will also require other must also be taken into consideration. integrated resource management activities Capital investment in irrigation, together with the O&M budget, makes up the lion’s share of the In Cambodia, the MOWRAM is in charge of budget allocations to MOWRAM. Focusing again sustainable water development, water resources on 2015 and 2016, the first two years of program management (surface water and groundwater), and budgeting implementation, irrigation capital investment delivery of water services, including irrigation for and irrigation scheme O&M amounted to around 80 agriculture. MOWRAM is responsible for implementing percent of total expenditure allocations (Figure 54). the Law on Water Resources Management of 2007 Planned allocations jumped in 2016 because O&M and related regulations. Some of the core functions spending was only introduced in the BSPs in 2016. foreseen in the law, as well as in the RGC Rectangular The payroll represented only around 4.5 percent of total Strategy Phase III and the NSDP 2014-2018, include: MOWRAM allocations in 2015 and 8 percent in 2016. to develop, expand, and rehabilitate irrigation systems that meet quality design and construction standards; to In recent years, government-funded capital expand water resources monitoring and forecasting; to investment has gone mainly to the rehabilitation of enhance flood prevention; to establish a national water irrigation schemes, while new scheme development resources inventory; to develop regulations for river Figure 54. Capital investment in irrigation and O&M make up the largest share of MOWRAM budget allocations Budget Strategic Plans and actual allocations to MOWRAM 100 80 60 USD million 40 20 0 2015 Planned 2015 Allocated 2016 Planned 2016 Allocated Irrigation capital investment Wages Irrigation scheme O&M Other activities (sub-program 1.2) (sub-program 5.3) (sub-program 1.1) Source: MOWRAM, BSP 2015-17 and 2016-18, MEF Budget Law 2015, MOWRAM Annual Work Plan 2016. 110 Cambodia: A Public Expenditure Review Figure 55. The significant boost in new irrigation schemes has been funded by DPs, particularly China DP and government funding to irrigation Funding by Development Partner 120 100 90 100 80 70 80 60 USD million 60 50 40 40 30 20 20 10 0 0 2010 2011 2012 2013 2014 2015 ADB AFD Australia China Japan JICA KOICA Donor-new irrigation schemes Government-O&M 2010 2015 Donor-rehabillition Government-rehabilitation of existing schemes of existing schemes Source: RGC investment: MEF Treasury Reports for 2010-2014, Budget for 2015; DP investment: CDC Database [accessed June-December 2016] and consultations. has been undertaken by DPs. Investment in new irrigation schemes and 50,000 ha from rehabilitated irrigation schemes increased from USD 23 million in irrigation schemes). As of 2016, the allocated USD 11.9 2010 to USD 114 million in 2012, driven by DPs (Figure million O&M budget could support approximately 26 55, left panel). Among DPs, Chinese investment projects percent (258,700 ha) of the estimated annual irrigation make up the majority of the increased investment over area in Cambodia (1,000,000 ha for wet season, 2010 to 2015, followed by smaller increases by Korea, 300,000 ha for dry season).70 Notably, only one-third ADB, Australia and JICA (Figure 55, right panel). Since of the existing 2,700 irrigation schemes are considered many of the existing irrigation schemes were damaged to be fully operational. Further analysis is needed to (e.g., by flooding) or not fully functional, DP funding for determine which of the existing schemes are suitable the rehabilitation of existing schemes has also increased for O&M support, rehabilitation and O&M support, or to over the past few years. Likewise, government capital be written off. Evidence suggests that with no O&M, the investment to support the rehabilitation of existing internal rate of return of an existing scheme drops below irrigation schemes has increased, rising from USD 12 percent after about 17-25 years (Egis, 2014). This 30 million in 2010 to USD 35 million in 2015. At this finding highlights the importance of O&M allocations, stage, there is no government budget to support the especially at a time when large investments in new construction of new irrigation schemes. schemes are being made. The recent increase in budget allocations to Integrated water resource management (IWRM) O&M is a positive development, but is likely to be activities, which are key to ensuring the sustainability insufficient for covering the entire irrigated area. of water availability, also seem to be underfunded. Although definitive figures do not exist, total capital The budget allocated to supporting other IWRM investment is estimated to support the establishment of activities has increased significantly in recent years, 200,000 ha of new irrigation area (150,000 ha from new from USD 5.3 million in 2015 to USD 9 million in 2016 70  Based on an O&M rate of USD 46 per ha as calculated under the ADB Water Resources Management Sector Development Program – Kong- shot O&M and completion cost estimate report. Cambodia: A Public Expenditure Review 111 (see Figure 54 above, ‘other activities’). If equally divided in place for managing water usage, it will difficult to among the 25 sub-programs included there, the average achieve sustainable development of Cambodia’s water allocation would be USD 360,000. In practice, however, resources, and investors’ confidence is likely to decline. not all sub-programs received budget allocations. For Recent water resources assessments have shown that example, only three of the six sub-programs under some river basins in the northwest of Cambodia are water resources and meteorological information already stressed.71 management (Program 4) received a funding allocation, and the total allocation for the program was only USD Increased funding to O&M and IWRM activities 16,500. This amount is unlikely to be sufficient to cover would require adequate staffing. MOWRAM counts the water resources monitoring needs of Cambodia’s 39 around 1,250 staff, of which 620 are based in the river basins, although it should be noted that, allocations national office departments and 630 across 25 provincial to Program 4 were adjusted to USD 238,000 in 2015, departments (Egis, 2012). At the national level, almost according the 2016 program budgeting document. 40 percent of the staff is located in the Department of Engineering, representing over four times the number Furthermore, there are no sub-programs to of staff in all other departments (Figure 56). In contrast, specifically support river basin planning and water the Department of Water Resources Management allocation planning, the strengthening of water and Conservation, which is in charge of most IWRM resources data collection, and establishment of planning responsibilities, has less than 5 percent of the a water resources modeling platform to support staff. With an average of 25 staff, provincial offices are water allocation and the setting of minimum flows. generally understaffed, making it difficult for them to Without a credible, science-based understanding make progress in establishing management systems for of water availability and without transparent rules water resources, including O&M of irrigation schemes.72 Figure 56. The highest proportion of MOWRAM staff is in the Department of Engineering Staff distribution across central departments of MOWRAM Technical service centre Internal audit Administration and human resources Finance Planning and international cooperation Farmer water user committee Water supply and sanitation Engineering Irrigation and agriculture Meteorology Hydrology and river works Water resources management and conservation 0 50 100 150 200 250 300 Number of staff Source: ADB, 2012. Training Needs Assessment Report: Ministry of Water Resources and Meteorology. Prepared by Egis-eau. Phnom Penh. 71  In 2014, a water balance assessment of the Stung Sreng River Basin found that the river basin has potentially been under severe stress over February to May and may/will suffer a water deficit in April by 2025. Over USD 115 million of investment has taken place since the mid-2000s for construction of a large irrigation scheme and rehabilitation of a number of existing schemes in that basin. 72  A recent assessment of staffing requirements, under the ADB Water Resources Management Development Sector Program (CDTA) identified that the newly built 40,000-ha Kanghot Irrigation Scheme would require a minimum of 62 staff to support O&M of the scheme. This includes a scheme manager, operation and maintenance supervisors, local operators and labor teams. 112 Cambodia: A Public Expenditure Review Based on recent estimates, the Siem Reap provincial Program 2 (flood and drought management), but no department of MOWRAM, responsible for 20 percent funding has been allocated to Program 4 (Figure 57). (180,000 ha) of the irrigated area in Cambodia, would Support to Program 3 (enactment of laws and regulations require a workforce of at least 200 staff to fully support related to water) has also been scanty. As of 2016, only O&M of the irrigation schemes (compared to actual two of the four key sub-decrees aimed at supporting the staffing of 40). 73 This calls for a more detailed analysis 2007 Water Law had been passed, and few provisions of staffing requirements74 to support O&M of irrigation had been implemented. This includes the establishment schemes and water resource monitoring and reporting of a national water resources inventory, which is needed activities, in order to help ensure effective use of funding. to manage the development of Cambodia’s water resources sustainably. Cambodia could benefit from an integrated framework for efficient and National irrigation and drainage design and effective irrigation scheme development construction standards as well as the associated auditing systems are not yet established to ensure Most DP support has not yet resulted in obvious a minimum standard of design and construction. improvements in MOWRAM’s capacity to oversee In general, irrigation project feasibility assessments the sustainable development of Cambodia’s have not included a comprehensive analysis of current water resources. DPs have mainly been funding and projected water demands in the basin and rainfall activities corresponding to Program 1 (water resources projections to determine the level of reliability of water management and development of irrigation systems) in supply to the site. Project design and construction the MOWRAM budget75 as well as to a lesser extent assessment methods can also vary among DPs. Figure 57. DP support has been concentrated in Program 1 of the MOWRAM budget 700 606 Program 1: Water resources management 600 and development of irrigation systems 500 Program 2: flood and drought management USD million 400 Program 3: enactment of laws and regulations related to water 300 Program 4: water resources and meteorological information management 200 Program 5: Administration and human 100 resources development 33 0.05 0.8 0 0 Program 1 Program 2 Program 3 Program 4 Program 5 Source: Council for the Development Cambodia Database, supplemented by consultations. 73  Reallocation of staff from the Department of Engineering to provincial offices may help address the current shortage of provincial staff over the next five years, although the reallocated staff must have the appropriate skill sets required at the provincial level. The reallocation could be offset by outsourcing aspects of their current responsibilities, such as the design of rehabilitation works of existing irrigation schemes funded by government. The Department would still have the oversight role to ensure designs meet relevant quality standards. 74  An organizational assessment of MOWRAM’s departments and PDWRAMs, in line with their objectives, would be needed to understand staffing constraints and areas requiring immediate support. This includes a review of actual permanent staff numbers and attendance rates and an analysis of the reasons for any low attendance. It should also include an assessment of the general technical capacity of the department and individual staff. 75  A few projects that have been accounted for under Program 1, such as the Australian Government funded Cambodian Agricultural Value Added Chain (CAVAC) project, the ADB Water Resources Sector Development Program (WRSDP), and JICA’s River Basin Water Utilization Water Program, have multiple objectives that include water infrastructure works and rehabilitation and IWRM activities such as strengthening water resources-related legislation, river basin planning, and supporting the establishment of self-sustaining FWUCs. The CAVAC program has had success in rehabilitating and establishing around 20 FWUCs to manage small (<500 ha) irrigation schemes in the southeast of Cambodia. Cambodia: A Public Expenditure Review 113 Different formulas have been applied for estimating existing schemes versus the construction of new ones, water requirements and reservoir capacity limits, and optimal size of irrigation schemes, security of water geographical aspects are not always appropriately supply, soil fertility, access to markets, and comparative considered. Development proposals may also fail to advantage crops. consider the economic, social, and environmental impacts across the basin, and some impact The uniqueness of water infrastructure, together assessments are not made public. In a number of recent with limited access to data, prevents the cases, irrigation projects have finished but have left a undertaking of an informative comparative cost scheme that is not fully functional due to poor design, analysis of new and rehabilitated canals and selection of material, and quality of construction. 76 If reservoirs. Not having national irrigation and drainage not addressed, it could affect the number of possible design and construction standards makes it difficult to crops per year and damage the canal structure, leading assess the quality of projects and to enable a level of to reduced returns to investment and increased O&M quality and consistency to be achieved and maintained. costs. For development of irrigation land, the average cost ranges from USD 2,000 to USD 3,900 per ha.77 It is important to undertake cost-benefit analysis Without knowing specific project breakdown costs, it for irrigation projects and assess project outcomes, is not clear what portion of total project funds have but only a few projects have undertaken such been allocated to a specific activity. The Cambodian analysis and have published the information. One Information System of Irrigation Schemes (CISIS) is of them is the CAVAC Phase 1 (2010-2015) project expected to help in understanding the condition of the that rehabilitated 20 small- to medium-sized irrigation 2,540 existing irrigation schemes across the country, schemes (up to 5000ha). Key findings were that most as well as the status of rehabilitation and construction schemes increased their crop cycles from two to three projects underway. However, CISIS data has not yet per year with an additional dry season crop, with the been made publicly available. following average yield increases: additional 6 t/ha per dry season, early wet crop increasing from 2.5 t/ha to 4.8 t/ha, and for wet season from 3.6 t/ha to 5.6t/ha. The average cost per ha was USD 2,148. The CAVAC The program budgeting experience program is unique among most DP projects in that it works directly with farmers, the private sector, and The roll-out of PB represents a key policy PDWRAMs and maintains their presence until the development aimed at eventually introducing irrigation scheme is fully operational and close to self- performance-informed allocations of public funding sustaining. More focus is now needed on understanding to increase value for money. PB falls under the broader the success of other projects and the potential costs Public Financial Management Reform and, following and benefits of project proposals. A starting point its pilot phase, was fully introduced in ten Ministries could be a nationwide assessment to identify suitable (including MAFF, MOWRAM, and MoEYS) in 2015. PB locations and activities that could deliver high returns to is expected to help shift the focus from activities being irrigation investment. A range of factors would need to determined by funding interests to policy guiding what be assessed, including the value of the rehabilitation of funding should support. Funding is directly linked to key 76  For example, recently completed secondary canals in Stung Sreng (25,000 ha) and Spean Sreng (4,751 ha) become completely submerged during the wet season due to poor drainage and lack of flood control structures. 77  This is a rough estimate based on limited information from six projects: ADB Water resources management sector development program, ADB Tonle Sap lowlands rural development project, ADB Northwest irrigation sector project, CAVAC, Kangshot irrigation development project (China fund- ed), and Stung Sheng Water resources development project (China funded). For large irrigation schemes (>5,000 ha), the cost of the development of irrigation land is likely to be higher as these estimates are unlikely to include all major costs such as full completion of tertiary and quaternary canal systems, necessary flood control works, and training of PDWRAMS to operate and maintain the scheme. 114 Cambodia: A Public Expenditure Review deliverables and measurable outputs, allowing for the is needed, especially in the areas of data collection and effectiveness of spending to be assessed. Ministries monitoring as described below. have to make trade-offs between available funds and reform priorities by weighing which activities are Overall, budget execution at MAFF has been essential now or can be postponed, reduced in scale, improving driven by personnel execution, while or canceled. All Ministries are to submit a five-year MOWRAM overall performance is dominated by Strategic Development Plan in line with the NSDP and variable capital execution rates. The MAFF budget annually submit a three-year rolling BSP and an Annual execution rate markedly improved from 85 percent in Work Plan to MEF. 2011 to an average of 94.5 percent during 2012-2016, pointing to increased absorptive capacity in the Ministry. Significant progress in PB has been achieved since A breakdown into wage and non-wage execution 2015. The MEF and relevant line ministries have been rates shows that the improvement was led by wage involved in a continuous process of improvement, expenditures, with non-wage execution first improving consolidating the number of sub-programs and refining then seeing a modest reduction (Figure 58, left panel). indicators year-to-year. MOWRAM is one of the few This pattern is also evident in MoEYS, another PB ministries that has included DP-financed capital into implementing ministry. The pattern could reflect general the Strategic Development Plan and BSP, although execution improvements, with some adjustment issues these have not yet been aligned at the sub-program in early PB implementation for non-wage execution, level. MAFF has also excelled in a number of PB areas, while the automatization of wage payments (introduced including setting out sub-programs already in the ASDP, in 2013) has improved wage execution significantly. assigning annual targets (from 2014-18) for all programs For MOWRAM, personnel execution has been solid and subprograms in the BSP, and including government throughout the period, but the overall execution rate payroll in PB documents.78 These developments lay the is determined largely by capital execution given its foundation for effective use of PB, although further work dominant share of expenditure (Figure 58, right panel). Figure 58. Budget execution has improved at MAFF, and personnel execution has been solid at MOWRAM, although overall execution largely is determined by capital execution Execution Rates, total, wage and non-wage, MAFF (left side) and MOWRAM (right side) 110% 140% 130% 100% 120% 110% 90% 100% 90% 80% 80% 70% 70% 2011 2012 2013 2014 2015 2016 2011 2012 2013 2014 2015 2016 Total Wage Non-wage Total Wage Non-wage Capital Source: MEF Treasury Accounts. 78  However, this practice was ended in the 2017 budget, as it was considered too cumbersome. While this is a step back with regard to out- come-based budgeting and accountability, it is significant that it has been done and can be reintroduced when considered appropriate. Cambodia: A Public Expenditure Review 115 Overall, the BSPs are aligned with sectoral and in the 2015 budget to 156 as of 2017. There is scope national strategies, although there is room for further for further improvement in selecting priorities at the improvement. For example, a few key mandates of the sub-program level when the annual work plans and the Water Law, such as the establishment of a national water budget strategic plans are submitted to the MEF. licensing framework, a water user fees framework, the implementation of protected water use zones and flood While indicators have been defined for all sub- retention areas, and the improvement of groundwater programs, not all of them are appropriate for management, do not seem to be covered in specific measuring progress toward goals. In some cases, sub-programs. A specific sub-program has been indicators are not SMART (specific, measurable, established for implementing the Sub-decree on Farmer assignable, realistic, and time-bound). For example, Water User Committees but not for the Sub-decree MOWRAM’s sub-program 1.1 (irrigation system on River Basin Management. In addition, the quality of maintenance and process) has a target of 331 improved these documents is not consistent across Ministries, schemes by 2016, but no explanation of how to with BSPs often being unrealistic due to the lack of an determine that an improved scheme was provided. In overall strategic budget framework, thus limiting the other cases, many sub-programs target outputs such role they play in decisions about the annual budget. as “staff trained, workshops organized, and meetings For example, MOWRAM BSP figures for 2016 (USD held,” but the outcomes of the training and meetings are 90 million) doubled those of 2015 and were 33 percent not specified and measured. While output indicators can higher than actual allocations. be appropriate and important for measuring execution, they are generally insufficient to measure performance The implementation of PB has followed international in terms of the effect—and thereby quality—of activities. good practices by limiting the number of programs For example, indicators for sub-program 1.8 (strengthen and goals, with ongoing efforts to consolidate and improve agricultural extension system and service) the number of sub-programs and activities. The under MAFF measure outputs (i.e. farmer outreach and number of programs has been limited to five per number of extension officers trained) but fall short of ministry, although there does not seem to be a clear measuring the extent of adoption of farm technologies selection criterion. In the case of MAFF or MoEYS, promoted by extension services and their links with programs seem to follow the administrative structure research. Overall, indicators associated with ASDP of the Ministry, whereas in MOWRAM, they are aligned Programs 2, 3, and 4 are generally sound, whereas with the strategic areas specified in the Rectangular others could be reformulated or complemented (Table Strategy. 79 The ASDP 2014-18 has 60 sub-programs, 33 proposes some additional indicators). compared to 28 in MOWRAM’s Strategic Development Plan. A large number of sub-programs complicates the A more daunting challenge is that data is not budget allocation process and program monitoring. For systematically collected for the indicators already example, in 2015, there were 6 sub-programs related defined due to the lack of an effective monitoring to rubber sector development, which could be merged system. As highlighted in a recent program budgeting into a smaller number of sub-programs since they review (World Bank, 2016), performance data is not share a similar goal. Some progress has been made available on a sufficiently complete, timely, or reliable in consolidating sub-programs and activities, with the basis to track performance, and there is no process in number of activity budget lines in MAFF falling from 415 place for regular monitoring and evaluation of programs. 79  MAFF: 1. Improved Agricultural Productivity, Diversification and Commercialization; 2. Animal Health and Production; 3. Sustainable Fisheries Resource Management; 4. Sustainable Forest Resources and Wildlife Management; and 5. Strengthening Institutions, Improving the Efficiency of Supporting Services and Developing Human Resource in the Agriculture Sector. MOWRAM: 1. Water resources management and development of irrigation systems; 2. Flood and drought management; 3. Enactment of laws and regulations related to water; 4. Water resources and meteorological information management; 5. Administration and human resources development. 116 Cambodia: A Public Expenditure Review Table 33. Additional indicators could be used to help measure the impacts of ASDP Programs Programs Existing indicators Proposed additional indicators Program 1: Improved agric. - Areas under all crops - Cropping intensity of land productivity, diversification, - Paddy yield - Yield of non-paddy crops and commercialization - Paddy surplus - Area of paddy applying GAP/SRI systems - Number of farmers adopting technologies promoted by the extension services - Improved varieties (by crop) released by research Program 2: Support to ani- - Livestock production - Number of farmers reached by the animal mal production and health - Ratio of animals vaccinated promotion extension services - Slaughterhouses imple- - Number of farmers applying good animal mented good health prac- husbandry practices tices - Number of farmers adopting bio-digester technology Program 3: Sustainable fish- - Aquaculture production - Number of operational fishery communities eries resource management - Capture fisheries production - Number of fishermen satisfied with exten- - Rate of processors that sion services improved quality and safety - Fish seed and fingerlings produced of fishery products Program 4: Sustainable - Protected and wildlife con- No need for additional indicators forest resources and wildlife servation areas management - Reforestation - Forestry communities es- tablished and strengthened Program 5: Support to - Ratio of activities achieved - Employment rate of graduated students institutions, support services, in the PB - Farm input shops inspected and in compli- and human resources - Students graduated ance with the law - Staff obtained local and - Number of women in management positions overseas training in MAFF Source: ASDP 2014-2018, World Bank staff illustrative suggestions. This limits the usefulness of program budgeting, since and irrigation, but this is not reflected in program lack of performance information prevents improvement budgeting documents. For example, for subprogram in the allocation and effectiveness of public resources. 1.8 (strengthen and improve agricultural extension In this PER, lack of access to data has been a severe services), the government allocation in 2015 was KHR limitation, precluding the analysis from linking inputs 2.2 billion, while the KHR 52.5 billion allocated by DPs (funding) with program outcomes. to agricultural extension was not reflected in the BSP. Meanwhile, government payroll is managed centrally, DP funding and public payroll are often excluded and the information for mapping personnel to specific from the BSP, which prevents analysts from sub-programs is insufficient. Without explicit accounting determining total resources devoted to the different for DP funding and the payroll within program budgeting, programs and activities. DP funding comprises about it is not possible to properly attribute the achievement of two-thirds of total public expenditure in agriculture activity outcomes. Cambodia: A Public Expenditure Review 117 Policy options to support and irrigation rather than to spend more, which will agricultural diversification and require significant improvements in coordination resilience among actors. As noted earlier, Government spending on agriculture and irrigation is now similar to the level observed in other middle-income economies, and if Going forward, Cambodia may not be able to rely on DP spending in agriculture is also taken into account, the same factors that drove agricultural growth in the total public spending (1.45 percent of GDP in 2015) past. Agricultural commodity prices are not expected to would be higher than in most countries. At the same return to the levels that prevailed in the late 2000s to time, while there are examples of both government- and early 2010s. Setting aside forested and protected areas, DP-supported projects that have had positive effects the scope for agricultural land use expansion is also on farm productivity and income, the impacts have much more limited now than in the past. In addition, tended to be localized. This calls for more proactive Cambodia is expected to find it increasingly difficult to and strategic engagement among public actors and compete internationally, regionally, and locally on the DPs during project selection, design, and preparation basis of low-cost, high-volume production. to ensure strong alignment with national priorities, coordination and linkages with other ongoing activities, New drivers of inclusive growth will need to be and sustainability of efforts. The next agricultural plan nurtured through the provision of core public goods. could include a vision for irrigation, and efforts should Going forward, Cambodia’s agricultural sector will need be made to prioritize and integrate potential DP pipeline to generate ‘more from less’—that is, realize higher projects into budget planning. The Technical Working levels of farmer, fisherman, and consumer welfare and Group on Agriculture and Water (TWGAW) would non-farm spillovers, while improving its natural resource be the natural forum for discussion and will require use efficiency and reducing its overall environmental significant revamping of its agility and decision-making footprint. To achieve this, a gradual transformation of responsibilities. the sector would involve a combination of increased irrigation, more diversified production systems, and There are multiple opportunities to improve PB and its greater value addition, which would contribute to links to sectoral performance. Priority attention should increased productivity (of labor, land, and water), be given to: (i) further consolidating sub-programs to greater climate resilience, and higher farm income and facilitate their management and increase synergies, while supply chain employment opportunities.80 This will at the same time improving their alignment with higher- require improving the availability of core public goods level goals;81 (ii) detailing, in sub-program work plans, while strategically aiming to differentiate Cambodian their general background and objectives, key activities, agriculture (and agro-industry) on a qualitative basis. outputs, timelines, and staffing requirements (skill sets This section provides some policy options aimed at and number of staff) for both national and provincial orienting public spending toward these broader goals departments; (iii) developing and applying lower-level (Table 34). and more attributable outcome indicators, adhering to the SMART principles; (iv) improving data collection Except for a few specific areas highlighted below, to better track the performance of sub-programs; (v) the focus going forward should be to improve the where possible, piloting then improving upon outcome- allocation and execution of spending in agriculture 80  Going forward, we would expect to see far fewer people regularly employed in rice farming, more people engaged in horticultural and livestock production, and increasing numbers of people involved in upstream (i.e. mechanical services) and downstream (i.e. food processing; food service) dimensions of the agro-food system. 81  Core IWRM areas of focus include: river basin planning and water allocation planning; strengthening of the hydro-met monitoring equipment and corresponding O&M program; and establishment, maintenance, and operation of a water resources modeling platform to support allocation of water and minimization of water impacts. 118 Cambodia: A Public Expenditure Review Table 34. Policy options to support agriculture diversification and resilience Challenge Short-term policy options Medium-term policy options (1-2 years) (3+ years) Value for money and impact Enhance strategic engagement and coordi- of public spending in agricul- nation among sector actors by revamping ture and irrigation has so far the TWGAW been limited Continue consolidating the number of BSP sub-programs and improving alignment with national priorities, while systematically collecting and monitoring performance data There seems to be underin- Undertake an institutional assessment of Elaborate a national agricultural vestment in a series of core current capabilities and future resource research strategy addressing technical and regulatory needs in inspection services, plant and research priorities as well as insti- functions (public goods) animal health, and water usage tutional and financing matters Collaborate with the private sector to de- Improve the efficiency of water velop business extension and/or incubator use in agriculture from off-farm models to support agribusiness SMEs and improvements with technology facilitate value chain integration transfer for on-farm investments Only one-third of irrigation Focus on rehabilitating existing irrigation Take into account both water schemes are fully functional schemes that have a secure water supply, availability and farm profitability and providing capacity building and funding considerations for new invest- to operate and maintain rehabilitated ments, and adopt national con- schemes struction standards Undertake regular project unit-cost analy- sis, combining the information with that in the CISIS (on condition of schemes), and use it to support informed decision-making Source: World Bank staff elaboration. based budgeting and performance measurement; and distributors and tourist facility operators will want (vi) including DP funds in PB documents, at least for to ensure that Cambodia has strong capabilities in information, to justify the government budget allocations pest and disease surveillance, quarantine services, for subprograms and explain the outcomes achieved.82 laboratory testing for contaminants and residues, and other areas. In the case of MOWRAM, adequate river To ensure Cambodia’s future success in diversifying basin assessment and planning, water licensing, and and adding value to its agriculture, increased groundwater management are essential to guaranteeing resources will be needed for a variety of technical the sustainability of both natural resources and returns and regulatory functions. For example, improved to investment. While increased public spending in these capacities will be needed for a broad array of sanitary areas is likely to be needed, this should be considered and phytosanitary service functions, both for the as part of a broader institutional assessment of domestic market and for supporting exports. Trading current capabilities and future needs (both in terms of partners, citizens, as well as modern food retail funding and staffing). The same would apply to other 82  The preparation of Annual Operation Plans for the MoEYS, while not complete, provides useful guidance for this. DP funding is included in sep- arate columns, making it easier to identify total sub-program allocations and the interplay between DP- and government-funded projects. Cambodia: A Public Expenditure Review 119 types of regulatory services, for example in relation to Additional resources might also be needed to upgrade or agriculture input quality as well as fisheries, forestry, create supplemental programs for agricultural education and biodiversity protection. Not all of these functions and vocational training. In addition, the efficiency of water will need to be performed by government agencies— use in agriculture could be increased by supporting indeed, the primary responsibility for food safety lies technology transfer for on-farm investments. with food providers, with the government providing the rules, oversight, and in some cases technical and other Over the next few years, government funding support for the adoption of improved practices. should continue to focus on the rehabilitation and sustainability of existing irrigation schemes rather Cambodia and its technical partners should chart than on the development of new schemes. First, out a national agricultural research strategy, akin given resource limitations, authorities could start by to the one recently developed for agricultural determining a set of priority regions at the river basin extension. Agricultural innovation has played only a scale, based on water availability, soil fertility, and limited role in past agricultural growth, but this will be access to agricultural markets. Second, government- increasingly important in the future as the opportunity funded capital expenditure should focus on the for further land expansion closes. Greater investment in rehabilitation of small- (<500 ha) to medium-scale (500- agricultural research will be needed to underpin the drive 1000 ha) irrigation schemes, with those that are proven toward a more diversified and resilient agricultural sector. non-viable being retired and removed from the asset Cambodia has so far invested only modestly in agricultural registry (CISIS). Third, O&M budget and staffing could research, although the recent spending increments are be increased in line with the size of the command area welcome. A national agricultural research strategy could under rehabilitated schemes. address not only long-standing challenges but also emerging ones such as climate change, environmental To ensure the functionality and sustainability of water management, nutrition, and post-harvest management. infrastructure projects, minimum water security As with extension, the approach should be to strengthen requirements and construction standards need to a pluralistic system of public, private, and other actors be developed and integrated into future feasibility and to deploy institutional arrangements that will ensure assessments and contracts. To date, feasibility studies that the increased agricultural research is demand- have focused mostly on design aspects and little on driven and both technically sound and cost-effective. water security and agricultural extension considerations Both local and international institutional partnerships that are essential for minimizing investment wastage could be promoted, while consideration could be given and risks (including potentially adverse economic, to the development of multi-stakeholder councils, social, and environmental impacts). A set of minimum competitive research grant schemes, and other types of requirements related to water security (Box 12) should mechanisms that have improved agricultural innovation be established and embedded in the feasibility study system performance in other countries. and should play a critical role in project selection. Notably, more investment in non-structural activities Increased investment may be needed to effectively will be needed to enable credible assessments of water develop the next cadre of entrepreneurial farmers security for project proposals, including: improving and agro-food business operators to help ensure understanding of current and potential water availability; inclusive growth in the sector. While the provision of managing scarce water resources at the basin scale; advisory and other services to SME agro-enterprises minimizing water pollution; and issuing water rights to is often not done by government agencies, there may allow farmers, water authorities, and other water users be a rationale for the government to co-finance certain to more confidently plan, invest, and progress. For each support programs of business extension or incubators. new irrigation scheme being developed, alternative land 120 Cambodia: A Public Expenditure Review use options and the agricultural service implications (PPPs) and other forms of collaboration. Elsewhere, of these choices should be considered. In addition, PPPs have been applied in relation to irrigation, market, national irrigation and drainage design and construction and other rural infrastructure, as well as several other standards should be prepared in line with international areas relevant to Cambodia’s rural development. Close best practices, and both government and DP-funded collaborative efforts can be used to promote the transfer irrigation investment projects should be required to of technologies, strengthen farmer organizations, meet these standards before proceeding. manage biosecurity risks, and address nutritional concerns, among others. As the sector matures, the In addition, to ensure return to investment in government may need to play less of a leadership and irrigation, introducing considerations of crop service provision role while taking on a widening set of profitability would be necessary. So far, the bulk facilitative functions. This transition in core functions of irrigation systems remain dedicated to rice mono- will have implications for public spending and, of equal cropping. Going forward, irrigation decisions need to importance, the range of technical and people-related be informed by considerations of farm profitability, the skills required of MAFF staff. These (human) resource changing domestic food demand, and Cambodia’s implications should be considered as part of a deeper competitive prospects for agri-food exports in the institutional analysis of MAFF functions and in the medium and long terms. context of the development of the next strategic plan for the sector. Assessing and enhancing the effectiveness of water infrastructure investments would require systematic unit cost assessments and collection of information. Undertaking regular unit cost assessments of ongoing and planned DP- and government-funded projects would help identify cost-saving measures. This could lead to a national assessment of the unit cost of water infrastructure projects to be undertaken with national irrigation and drainage design and construction standards as a benchmark. The information on unit costs could be combined with CISIS, an important tool for monitoring the condition of irrigation schemes. As a next step, it would be important to review existing data collection, storage, and reporting procedures to improve the quality and credibility of the information in the dataset to support informed decision-making. For the CISIS tool to be useful, access could be granted to all MOWRAM staff and DPs working on irrigation scheme projects. Over the medium to long term, CISIS should ultimately become a primary data source for supporting the prioritization, development, and assessment of irrigation project proposals. Finally, it should be noted that delivery of public goods should not be based exclusively on public funding, as there is room for public-private partnerships Cambodia: A Public Expenditure Review 121 Box 12 Water security framework AVAILABILITY, ACCESS, LEGAL SECURITY RELIABLE WATER & RELIABILITY OF WATER SUPPLY DELIVERY SYSTEMS OF WATER SUPPLY AVAILABILITY, ACCESS, & RELIABILITY OF WATER SUPPLY • Basin-wide annual and dry season assessments of available water resources are required to determine how much water is available for use by new projects. Historical rainfall and recharge records, climate change projections, and existing water demands all need to be considered. • Available water for commercial use is to be within the sustainable water use limit set under the (basin-wide) water allocation plan, where one exists. This limit is set to give highest priority to the protection of human and aquatic ecosystem water requirements. • The extent of access to available water must be assessed. There could be abundant water resources but limited delivery capacity due to canal and pipe size or no delivery system in place. • All water users will want to know how often they are likely to receive a specific volume of water (‘reliability’ of water supply) to help them understand risks to water supply and plan ahead with confidence. • Crucial to determining availability, access, and reliability of water supply is the quality of water resources data (e.g., length, extent, and type of data records) and water resources modeling capacity to undertake these assessments. LEGAL SECURITY OF WATER SUPPLY • Water users not only need access to a reliable supply of available water but must also have a legal right to use it. A water right (e.g., license, permit, certificate) is a statutory instrument that sets out the total volume of water available for use, when, where, and by whom it can be used. • Water rights provide water users and investors with a level of certainty that they will regularly receive a specific volume of water, giving them more confidence to make more informed business decisions. • Effective water rights must be granted (issued) in line with basin-wide sustainable water use limits and existing water rights to avoid over-allocation of available water that can lead to costly government compensation. • An effective water rights system requires ongoing monitoring and regular reporting of water use to ensure water users are not taking more water than the amount to which they are legally entitled. RELIABLE WATER DELIVERY SYSTEMS • National water development design & construction standards should be established and followed to ensure a consistent, minimum level of development quality. • O&M procedures (including roles and responsibilities and available O&M funds) for new water infrastructure should be established within the first few years of project completion. 122 Cambodia: A Public Expenditure Review 6 IMPROVING THE VALUE FOR MONEY IN ROAD SPENDING Introduction and motivation countries, Cambodia still lags other structural peers in C terms of quality of transportation infrastructure, which does not seem to have significantly improved over the ambodia has achieved remarkable past six years (Figure 59, left panel). Cambodia ranked economic development, while 73rd in the world in 2016 for the Logistics Performance transportation infrastructure does not Index (LPI), and is below the regional average in the seem to have kept pace. Cambodia has attained transportation dimension. strong economic growth over the past two decades, driven by agriculture, garment exports, tourism and, Transportation in Cambodia predominantly takes more recently construction. Although total public place by road, while accessibility challenges investment has been substantially higher (an average of persist in some rural and remote areas. 7.5 percent of GDP in 2000-15) than in most ASEAN Transportation is predominantly provided by roads, Figure 59. Cambodia still lags behind its peers in the quality of transportation infrastructure Transportation infrastructure quality index Share of paved roads, 2017 7 100% 90% 6 80% 5 70% 60% 4 50% 3 40% 2 30% 17% 20% 1 10% 0 0% Guatemala Nicaragua Bangladesh Cambodia Philippines Thailand Philippines Thailand Malaysia Indonesia Cambodia Lao PDR Vietnam Vietnam Transport 2017 Transport 2010 Source: World Economic Forum, Global Competitiveness Indicators; ASEAN statistics. Cambodia: A Public Expenditure Review 123 and transportation services are dominated by implies changes in public investment management, motorcycles and 2-axles vehicles (together, about 90 public-private partnership mechanisms, and assets percent of the traffic). Modern forms of transportation management. Forth, improving institutional capacity, and road network systems are limited to main cities, planning and implementation, as shortcomings in terms connection with neighboring countries and with of organizational structure and division of responsibilities, key trade gateways, while the trucking industry is human resource capacity, and internal control systems fragmented and rural connectivity in many cases are acknowledged. remains informal and precarious. With 79 percent of the population living in rural areas as of 2016, an This chapter draws from the guiding principles set urban-rural gap in terms of access to services and in these national level strategies, and focuses on road rural accessibility remains. Cambodia’s road identifying ways to improve spending efficiency network has significantly increased in recent years, and transportation service provision in Cambodia. with total length increasing from 46,245km in 2013 The analysis presented in this chapter focuses in the to 61,379km in 2017. With the country’s strategic road sector, in the understanding that it is, by far, the direction of reaching out to rural population, about dominant transport mode of the country for people 75 percent of the network is comprised by rural and freight, as well as for domestic and international roads, mostly unpaved. This explains why only 17 connectivity. Section 6.2 discusses the basic foundations percent of roads were paved as of 2017 (Figure 59, of the road sector in Cambodia – legal and institutional right panel), as authorities have prioritized rural road context, sector strategy, and public spending process development, at the same time that focused paving and procedures; section 6.3 analyzes road spending efforts in the national road network first. This strategy patterns and trends, as well as road conditions; and is appropriate for Cambodia, given its income level section 6.4 presents policy recommendations to help and large investment needs. address the challenges identified. Under the current Rectangular Strategy, transport sector development continues to focus mainly on roads, with aspiration towards reaching The road sector in Cambodia— multi-modality, and emphasis in connectivity Institutional framework with neighbors and fostering of private sector participation. The Rectangular Strategy phase III The institutional framework for road identifies four main priorities to tackle transport sector construction, maintenance, and challenges. First, driving drown transportation costs, transportation regulation is fragmented by focusing in improving road repair and maintenance, among multiple agencies traffic safety, and traffic congestion. The second priority aims at developing multi-modal transport connectivity In Cambodia, the legal framework is provided by to ASEAN (especially Mekong basin countries), and the Law on Road and the Law on Land Traffic, both requires from developing the regulatory environment from 2014. The legal framework governing roads in and cross-border facilities, among other measures. Cambodia comprises the Road Law83, that focuses Third, increasing spending efficiency by shifting from on the physical aspects of the network development low-cost, high-maintenance needed infrastructure and maintenance, and the Law on Land Traffic84 that projects towards projects with higher standards; this structures road use, safety and signaling. 83  Law on Road, Preah Reach Kram No. NS/RKM/0514/008, adopted by the National Assembly in April 2014 84  Law on Land Traffic, Preah Reach Kram No. NS/RKM/0115/001, adopted by the National Assembly in December 2014 and became effective in January 2015 124 Cambodia: A Public Expenditure Review The Road Law designates the road type and roads is carried out directly by the Heavy Equipment classification and distributes road network Department at MPWT. Urban roads in Phnom Penh management responsibilities among the Ministry of are under the responsibility of the City Hall, while Public Works and Transport (MPWT), the Ministry of other urban centers are considered as provinces. In Rural Development (MRD), and local governments. terms of transportation services, the Department of The network is divided into six categories of roads: Land transport at MPWT is in charge of enforcing (i) expressways, (ii) national, (iii) provincial, (iv) rural, rules, while secondary regulation is not sufficiency (v) urban, and (vi) special roads. The Road Law also developed (e.g. with regards to overload control). The attributes jurisdictional responsibilities for development Phnom Penh Bus Authority (PTMA) is the only publicly- and maintenance of the physical network to owned urban transportation company, since services governmental agencies. As such, the Law designates have been traditionally provided by tuk-tuk. Most of the Ministry of Public Works and Transport (MPWT) as operators on the roads are private (trucking services, responsible for the development and maintenance of private vehicles, and non-motorized transport). expressways, national and provincial roads; the Ministry of Rural Development (MRD) is responsible for rural Maintenance of rural roads is significantly roads; cities and municipalities are responsible for urban fragmented. Rural road management is split between roads (Table 35). The management of special roads the MRD (for planning, construction, rehabilitation, however, is determined case-by-case. As of today, there and emergency maintenance of all rural roads), and are no expressways in Cambodia. local governments (entrusted with periodic and routine maintenance). While some degree of decentralization is MPWT manages both national roads and desirable, the intricate and fragmented division of labor provincial roads, and is also in charge of enforcing and resource allocation often results in funding being transportation regulation. MPWT manages road spread too thinly (Figure 60). This especially affects construction centrally, while using own provincial the maintenance of roads in levels T2-T4, which often departments (DPWT), to carry out maintenance remain in poor condition, making it more difficult for rural activities. The construction, and rehabilitation national population to access economic opportunities that are roads and emergency maintenance of national well-connected by national road network. Figure 60. Resource allocation structure for rural roads in Cambodia Donors National Treasury MRD MOI Maintenance T1 Provinces Maintenance Appointed T2 Districts Investment Maintenance T3 Communes Elected Maintenance T4 Villages Source: World Bank staff elaboration in consultation with authorities. Cambodia: A Public Expenditure Review 125 Table 35. Cambodia’s Road Transport Infrastructure Institutional Mapping Functions --> Planning Construc- Rehabilita- Periodic Emergency Operation Regulation Policy tion tion & Routine Mainte- of Umbrella Mainte- nance Services nance National Roads MPWT IMCRR (NR) MPWT/ DPWT in the (by Heavy with MPWT (1-digit and DPs provinces Equipment MPWT 2-digit) Dept) Provincial Roads IMCRR (PR) MPWT/ DPWT in the DPWT in the with MPWT (3-digit and DPs provinces provinces Private MPWT 4-digit) trucking National Rural Roads (T1) IMCRR Provincial services Assembly MRD/ Dept of (Provincial town with MRD Govern- MRD (fragment- through DPs Land Trans- to District center) MRD ments ed and the NSDP port under informal) / Rural Roads (T2) IMCRR MPWT District MPWT & (District center to with MRD MRD MRD Private MRD Offices Commune) MRD vehicles Rural Roads (T3) IMCRR (Commune to with MRD MRD Communes MRD Village) MRD Rural Roads (T4) IMCRR (Village to Village with MRD MRD Villages MRD or to the field) MRD Phnom Penh Bus Urban Transport services / road by PTMA PPCH PPCH PPCH PPCH MPWT Dept of (roads in and ring (PP Bus roads around Land Trans- Authority) MPWT / Phnom Penh) port under PPCH MPWT / Other Urban Private PPCH Transport / road vehicles / DPWT DPWT DPWT DPWT DPWT non-mo- (Others: MRT, NMT, etc.) torized Agricultural Private MAFF / Roads MAFF / MAFF / MAFF / MAFF / trucking MAFF / MAFF / MOW- (in line with T4) MOWRAM MOWRAM MOWRAM MOWRAM Private MOWRAM MOWRAM RAM vehicles Source: World Bank staff elaboration in consultation with authorities. Legend IMCRR Inter-Ministerial Committee for Reconstruction and Repairs PTMA Public Transport Management Authority MPWT Ministry of Public Works and Transport PPCH Phnom Penh Capital Hall DPWT Departments of the MPWT in each Province MAFF Ministry of Agriculture Forestry and Fisheries MRD Ministry of Rural Development MOWRAM Ministry of Water Resources and Meteorology 126 Cambodia: A Public Expenditure Review An additional level of institutional fragmentation There is no transportation specific sector and complexity is added by having the DPs play strategy a prominent role in financing investment. The two key road agencies, MPWT and MRD, have a different The transportation sector in Cambodia does not structure, are independent from each other in both currently count with a specific strategy of master the planning and execution stages, and technical plan. Both Rectangular Strategy phase II and phase III knowledge and capacity is minimal. In addition, the DP- documents envisioned the elaboration of masters plans funded projects follow separate budget and planning for Transport Infrastructure Development or for Urban processes than those financed using government Infrastructure Development, but these have not been budget. Differences include criteria for project selection, prepared. Some recent progress has been achieved on level of project preparation, ability to finance multi-year the preparation of a National Logistics Master Plan (Box projects and, very evidently, the relative relevance of the 13). In practice, the road sector strategic and policy project in terms of the size vis-à-vis the portfolio. The guidelines remain those stipulated in the 5-year National implementation of DP projects follows in most cases Strategic Development Plan (NSDP). The NSDP agreed Standard Operation Procedures that differ structures key policies and actions for each ministry and from government procurement processes, and use agency, and sets forth the monitoring and evaluation international contractors. (M&E) framework for the planning cycle. However, the Box 13 A National Logistics Master Plan for Cambodia In November 2017, the RGC established the National Logistics Council by issuing the Organization and Functioning of National Council of Logistics Royal Decree 2017. The National Logistics Council (NLC) is tasked with leading the drafting of the National Logistics Mater Plan (NLMP) from the Government side and to coordinate needs and inputs from different agencies. JICA and the World Bank are playing a critical role in supporting the government on the preparation of the NLMP. The NLC will be chaired by a Deputy Prime Minister, with senior officials from MPWT and MEF as members. The NLMP is a key instrument to develop the structural transformation of the Cambodian economy with industrial sector envisioned in the Rectangular Strategy (RS III 2014-2018) as a key driver toward reaching middle-income status. The strategic vacuum is becoming evident in practice, deterring potential investors and reducing the potential benefits that comes from coordinated investments. The objective(s) of the first NLMP is to conceptualize an integrated and effective multi-modal transport and logistics system that connects major economic poles creating three structural economic corridors: Phnom Penh–Sihanoukville, Phnom Penh–Bavet and Phnom Penh–Poipet. The NLMP will be further developed logistics sub-sectors and propose reforms that integrate various mode of transport (such as roads, ports, railways, and waterways) together – including the integrated regulatory framework. The scope and timetable of the NLMP is to be determined, but the target is within 2018. Cambodia: A Public Expenditure Review 127 NSDP does not provide an explicit articulation of plans for more detail). In the case of government-funded and objectives among different types of road networks projects, annual planning discussions are conducted (i.e. national roads, rural roads and urban roads) or at the Inter-Ministerial Committee for Rehabilitation and among transport sub-sectors (i.e. among roads, ports, Repairs (IMCRR). The IMCRR is co-chaired by MPWT airports and logistics services). and MEF (for national, provincial and urban roads outside Phenom Penh)85 and MRD and MEF (for rural Budget planning and execution has roads), and comprises various technical and policy improved in recent years, but continues departments in the line ministries.86 Authorized by the to present challenges Prime Minister, IMCRR meets bi-annually (i) to establish the prioritization criteria that will be included in the year The introduction of program-based budgeting in IMCRR Memo and (ii) to screen and select the specific recent years has helped improve the allocation of projects warranted funds from the Treasury. recurrent spending. The Budget Strategic Plan (BSP) is a systematic attempt to perform multi-year planning, The planning process for government-funded consolidate the operation, maintenance and investment investment starts with a long-list of investment budgets regardless of the source of funding, and projects, which may have originated from align them with sector priorities by defining programs politicians, provincial departments, and subnational and subprograms. However, integration of capital authorities. That long list is screened to make sure all spending in the BSPs for MPWT, MRD, and Phnom projects included are under a size threshold and can Penh City Hall (PPCH) has so far been limited. MPWT be implemented within a year time. With these criteria, establishes internally a “financing working group” to plan the Government guarantees that projects that could be for recurrent expenditures. Meanwhile, MRD works in eligible to be funded by DPs —larger and multi-year coordination with their provincial branches and elected by nature— are explicitly excluded from the long-list of commune authorities to determine the needs for road projects considered to be financed by the Treasury. That repair and maintenance. In addition, commune councils long list of projects is then screened with the application are also allocated some funding for the maintenance of of yearly-revised prioritization criteria, formalized in the T3 & T4 rural roads, as well as for other infrastructure, IMRR prioritization memo. The screening of the long- such as small-scale irrigation. At the subnational level, list of projects leads to the preparation of a short-list each provincial government prepares a BSP on behalf of projects that is taken to the Prime Minister around of the district/municipality and commune/village level; November for its clearance (Figure 61). in this case, the annual planning process goes through the Ministry of Interior (MoI). This multi-actor and multi- There is a relatively fixed set of prioritization criteria layered process reflects the overall fragmentation of for government-funded projects, although some are road maintenance efforts in Cambodia. relatively open and its ranking is not clearly defined. Recurrent IMCRR filtering criteria include projects that Infrastructure investment projects are split into those are considered backbones of the economy, roads that that are government-funded and those DP-funded, were damaged by flooding or other factors, ongoing and both follow different rules and processes. projects that need to be finished to maximize economic Investment funded by DPs follows PIP discussions effectiveness, those related to protecting national (see chapter on Public Investment Management sovereignty, especially along the borders, and those 85  Since 2016, IMCRR also covers inland waterway and railways. MEF has provided about US$20 million per year for track rehabilitation during 2016-201 as a follow-on to ADB project. 86  As a way of example, for the IMCRR for national, provincial, and urban roads, it includes from the MPWT their Technical General Department, Public Works General Department, Road Department, Technical Research Center, and Policy Department. 128 Cambodia: A Public Expenditure Review Figure 61. MPWT and MRD Budget Planning Timeline Budget preparation starts May Jun Jul Aug Sep Oct Nov Dec Jan Feb IMCRR considering Ministerial Council National Assembly Projets for long-list Discussion and Appruval Approves the Budget (short-list) Source: World Bank staff elaboration in consultation with authorities. decided by Samdech Techo Prime Minister. There is also between the initial request and actual allocation can a mention that, <<[b]esides the above prioritization rules, reveal underspending, under-maintenance and a selection of projects shall be based on the principles formidable funding gap for roads. of equity between provinces, based on demographic factors, necessity to enhance development and In terms of funding for investment projects financed necessity due to political reason>>. However, the way by Treasury, allocations are determined annually the principles of equity and demography are applied is even if the project is multiannual in nature, which not clearly defined. makes its completion uncertain. A filter applied for investment project screening during the national budget Only between 5 and 6 percent of the long-list gets preparation process is that the project needs to be fully financed each year, which may signal weaknesses in executed within the fiscal year. Such criterion limits the the planning process, and retention of proposals not size of the investment projects to less than $10 million fully developed or that may be no longer relevant. In and explicitly excludes large infrastructure construction.87 2016, the implementation of the entire long-list of MPWT This $10 million limit would fund approximately 12km of projects would have cost approximately US$3 billion, concrete national road construction, or about 30 to 50km while the budget available for both construction and of asphalt concreate national road construction. The maintenance was US$160 million. Similarly, for the fiscal imposition of this limit forces implementing agencies to year 2018, the MRD long list included approximately artificially split multi-year projects in a sequence of annual 2,600 road maintenance requests from villages, while projects. It also creates uncertainty regarding securing MRD could accommodate only 165 requests due to funds for on-going multiyear projects, and creating delays budget constraints after the annual screening. Some in the provision of funding (as portions of the infrastructure projects in the long-list might be overly optimistic, need to be proposed and screened every year), as well unfeasible and only respond to political motivations. In as pauses in the implementation. This is likely to increase addition, the quality and readiness of project proposals the final cost (due cost and time overruns, and rebuild in the long-list may vary —some would include feasibility of previously constructed phases), and could result in studies; some are just concepts that represent certain white elephants (e.g. building the support of a bridge “wishes”. On the other hand, the large difference and then abandoning a half-finished unusable asset). 87  The only historical exception is the reconstruction of National Road 68 following border disputes with Thailand. That project amounted to $30M and was multiyear, 2009-10. MPWT essentially used its whole budget for the execution. Cambodia: A Public Expenditure Review 129 Figure 62. Budget Execution Timeline A. MPWT MPWT prepares the 1st Joint 2nd Joint “specification book” inspection of works by inspection of works by submitted to MEF MPWT & MEF (or IMCRR) MPWT & MEF (or IMCRR) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec MEF releases 1st MEF releases 2nd MEF releases 3rd installment of funds installment of funds installment of funds B. MRD MRD prepares the technical MRD Provincial Departments The last 10% designs and specification request for more fund of funding is submitted to MFD at 50% progress for defect liability 50% progress 1 month Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec MEF releases 40% of funding MEF releases 50% of funding to Provincial Departments approximately 1 month after requested Source: World Bank staff elaboration in consultation with authorities. In addition, a rigid budget execution timeline for the tendering process (including detailed design) for road projects results in significant overruns. The rural projects could take more than 6 months, with the fund release schedule is predictable but relatively rigid, start falling into the rainy season (Figure 62b). and traditionally results in delays. In the case of MPWT, following National Assembly approval of the budget, the Proposed investment projects to be financed by MPWT consolidates technical specification books from DPs enter the Public Investment Program and 25 provinces, which can vary significantly in quality, and follow different selection and implementation rules. the review process, taking place in February, may get The criteria by which a project is funded by DPs rather delayed. The first release of funds occurs around March, than by the Treasury does not respond to its complexity which leads inevitably to overlaps of works with the but rather to its size and whether it fulfils certain basic rainy season —which spans May to November (Figure construction standards. For MPWT, usually, projects 62a). Rainy season maintenance may cost about 40-50 considered for PIP are multi-year, reconstruction percent higher than that during the dry season. Similarly, projects that are larger than US$10 million each. 130 Cambodia: A Public Expenditure Review Meanwhile, for MRD, earth or laterite rural road 2016 (Figure 64, left panel). This has been partly offset improvement projects would use government funding, by increasing national budget allocations, from US$150 while road improvements with concrete or pavements million in 2013 to US$286 million in 2016. As a percent would be listed under the PIP to pursue DP funding. of GDP, national budget allocations have averaged 1.4 The PIP is consolidated and prioritized by the Ministry percent in 2014-16, subnational budget 0.5 percent of of Planning (MoP), and contains a pool of proposed GDP, and cities 0.35 percent of GDP (Figure 64, right investment projects to be funded exclusively by DPs. panel). In 2017, some recovery in DP-funded spending The completion of feasibility study is not a requirement has been observed. for a project to be included in PIP. MEF plays a role of gatekeeper by checking the consistency of projects The Ministry of Public Works and Transport is the under PIP against national priorities and the national main implementing agency, while funding executed debt strategy. One of the problems with the existing through the Ministry of Rural Development and the fragmentation is that a project listed in the PIP waiting Phnom Penh City Hall has increased significantly. for DP funding for several years may be more of a priority Funding implemented through the MPWT declined than other projects included in the long-list, but would between 2014 and 2016, as DP-funded projects never get government funding. executed through this agency decreased, prior to experiencing some rebound in 2017. Meanwhile, road spending implemented by MRD increased to an average of US$70 million in 2014-16, and that implemented Prudent spending with rising through the Phnom Penh City Hall to an average of pressures US$64 million in 2014-16, more than doubling the average for 2011-13 (Figure 65, left panel). During Over the past four years, there has 2014-16, despite the decline, MPWT still implemented been a significant shift in the sources of more than half of total transportation spending (Figure funding for the road sector 65, right panel). Road spending has traditionally received larger In recent years, capital spending has been on funding than education or health, while it has been decline, while allocations to operations and on decline in recent years. Road spending peaked maintenance (O&M) have increased. Driven by DP- in 2012, at around 4.3 percent of GDP, following a funded spending, capital expenditure declined from 3.4 significant scaling-up of infrastructure investment, percent of GDP during its 2012 peak, to 1.4 percent of mostly driven by Chinese funding. Since then, it has been GDP in 2016, prior to experiencing some recovery in steadily falling, to an estimate of 2.6 percent in 2016 2017 (Figure 66). Meanwhile, O&M spending increased (Figure 63, left panel). Government spending in roads from around 1 percent of GDP to 1.2 percent of GDP over as a percentage of the national budget has remained the same period. The Phnom Penh City Hall and local around 10-12 percent of the national budget (Figure 63, governments mostly perform maintenance function, right panel). In 2016, education overtook roads in terms while leaving infrastructure building to DPs or other of public spending, as it has been designated a priority government agencies. In 2016, 70 percent of MPWT sector. executed spending went to construction activities, and 30 percent to operations and maintenance; at MRD, the The decline in public road spending has been mostly proportion was 57/43. driven by a decrease in funding by DPs. In nominal terms, DP funding peaked in 2013, at around US$ 350 The shares of personnel expenditures for both million dollars, before declining to 90 million dollars in MPWT and MRD are relatively small. In 2016, wage Cambodia: A Public Expenditure Review 131 Figure 63. Public spending in roads has been on smooth decline it peaked in 2012 Total Road, Eduction, and Health Sector spending, Government own spending in road, education, % of GDP and health sector as % of National Budget 5.5 14.0 12.0 4.0 10.0 3.0 8.0 6.0 2.0 4.0 1.0 2.0 0.0 0.0 2011 2012 2013 2014 2015 2016 2011 2012 2013 2014 2015 2016 2017 Road Sector Education Sector Health Sector Road Budget Education Budget Health Budget Source: National Treasury database 2011-16, CDC DP database 2011-16, and sub-national budget 2011-16. Figure 64. A decline in DP funded road expenditures has been partly offset by rising national budget allocations Road spending by source of funding, US$ million Road spending by source of funding, % of GDP 800 5.0 700 4.5 4.0 600 3.5 500 3.0 400 2.5 300 2.0 1.5 200 1.0 100 0.5 0 0.0 2011 2012 2013 2014 2015 2016 2017 2011 2012 2013 2014 2015 2016 2017 City/Urban budget Donor City/Urban budget Donor National budget Sub-national budget National budget Sub-national budget Source: National Treasury database 2011-16 and sub-national budget 2011-16. represented around 8.1 percent of total MPWT spending MRD, the wage represented about 6.6 percent of total (approximately US$16 million), with non-wage recurrent spending in 2016, with non-wage recurrent expenditure expenditures of 37.8 percent and capital expenditures of 41.5 percent and capital expenditure of 51.9 percent. of 54.1 percent. From the US$16 million wage From the US$4.6 million wage expenditure for MRD, expenditure for MPWT, US$10.4 million is for operations US$4.5 million is for operations (69 percent of operations (89 percent of operations spending), and US$6.1 million spending), and US$0.2 million is for overheads (4 is for overheads (63 percent of overhead spending). For percent of overhead spending). 132 Cambodia: A Public Expenditure Review Figure 65. MPWT is the main implementing agency in roads Road sector spending by implementing instiution, Share of spending US$ million (annual average 2014-16) 800 Joint MPWT/MRD 4% 600 400 City Hall 11% 200 Local Govt 15% MPWT 58% 0 2011 2012 2013 2014 2015 2016 2017 Non-Transport 1% City Hall Joint MPWT/MRD MPWT MRD 12% MRD Non-Transport Local Govt Source: National Treasury database 2011-16, CDC DP database 2011-16, and sub-national budget 2011-16. Figure 66. The decline in capital expenditure in roads has been driven by DPs Transport Spending by source for Capital Road spending by economic classification, % of GDP Expenditures and O&M, % of GDP 5.0 2.0 4.0 1.5 3.0 1.0 2.0 0.5 1.0 0.0 2015 2016 2015 2016 2015 2016 2015 2016 2014 2014 2014 2014 0.0 2011 2012 2013 2014 2015 2016 2017 City/Urban Donor National Sub-national budget budget budget Capital Expenditure O&M Capital Expenditure O&M Source: National Treasury database 2011-16, CDC DP database 2011-16, and sub-national budget 2011-16. Capital spending continues to focus on years. As discussed in the previous section, large- national roads, and there seems to be scale capital investment is usually undertaken by DPs underinvestment in provinces outside the whom, over the past two decades, have been in charge main corridors of building or reconstructing the main arteries of the road network. The recent decline in DP-funding project Three quarters of total spending is directed to has driven down the share of national road spending national roads, despite a slight decline in recent over total spending, from 87 percent in 2013 down to Cambodia: A Public Expenditure Review 133 less than 73 percent in 2017 (Figure 67). The share of of road investment projects. Among DPs, in recent spending directed towards urban and rural roads has years, China has been a major player, with their support increased significantly in recent years. mostly focused on national road construction (Figure 68). Among the largest projects are the reconstruction Development partners have been playing a major and improvement of national roads No. 5 and No. 6, role in Cambodia’s road sector investment. As supported by Japan and China, respectively. Both projects abovementioned, DPs have been the main supporter span several provinces to the west of Phnom Penh. Figure 67. Three quarters of total capital spending is directed to national roads Road capital expenditure by type of roads (2011-17) 100% 90% 80% 70% 60% 50% 40% 80.3 82.0 78.8 77.9 76.9 73.3 72.8 30% 20% 10% 0% 2011 2012 2013 2014 2015 2016 2017 National Road RR: T1 RR: T2 RR: T3 & 4 Rural Road Urban Road Other Road Source: National Treasury database 2011-16, CDC DP database 2011-16, and sub-national budget 2011-16. Figure 68. China and Japan have led Development Partner capital spending in transportation in recent years DP-supported road investment project 2011-17 1,400 1,318 1,200 1,004 1,000 800 Millones 605 600 400 267 191 214 200 32 65 38 5 0 ADV China Japan Republic of Korea World Bank Approved Disbursed Source: CDC DP database 2011-17. 134 Cambodia: A Public Expenditure Review Figure 69. Sparsely populated provinces in the north and northeast of the country have not received DP funding Donor-supported road spending Donor-supported road spending per population 2013-2017 (USD Million) 2013-2017 (USD per population) USD Million USD/Population 0.00 355.80 0.00 478.34 Source: CDC Database (2011-2017). During the past 5 years, DP-supported road spending condition and drivable all-year around (or during most has been focusing on provinces that connect to of the year). In this context, MPWT conducts road neighboring countries. This is the case of Koh Kong condition survey almost every year for all 1-digit national (US$478 per capita), Banteay Meanchey (US$424 roads and for a set of samples 2-, 3-, and 4-digit roads. per capita), both with important border crossing with Two-thirds of the 1-digit national roads are in good or Thailand and, to a lesser extent, Battambang, Kampong fair condition, compared to 31 percent of provincial Chhnang, and Prey Veng, on the way from Phnom Penh roads and just 23 percent of the 2-digit national roads. to Bangkok and Ho Chi Minh City (Figure 69). Most of The situation for the rural network is hard to assess, the projects undertaken are national road expansions. since a road survey has not been conducted. Lack of Meanwhile, provinces in the north and northeast, which data also affects any assessment of the investment or are also among the poorest in the country did not maintenance needs that could be crucial information benefit from DP funded road investments during 2013- to ensure the efficiency of funding allocation during 17 (Preah Vihear, Stung Treng, Ratanakiri, Mondul Kiri). budget discussions. Anecdotal information points to the majority of rural roads being in poor condition. The recent emphasis on operations and maintenance is aimed at addressing road The sub-national budget to maintain all rural roads deterioration in the country is just slightly above the budget of the Phnom Penh City Hall. It is worth noting that national About 87 percent of the national and provincial roads and some provincial roads often see heavy roads were in good or fair condition in 2017 (Figure trucking traffic, and deteriorate more rapidly; DPs focus 70). Aside from materials and quality standards of recent solely in the building and reconstruction of these main road levels, in relation to the expected traffic, what is arteries, with no maintenance funding associated. Thus, considered relevant when assessing connectivity under maintenance is entirely in the hands of the Cambodian a transportation network is whether the roads in good agencies (Figure 71), and budgeting and spending in Cambodia: A Public Expenditure Review 135 operations and maintenance has not always kept pace from sub-national budgets. Meanwhile, communes and with capital spending. An average of 46.9 percent of villages (elected), implement 2.3 percent of total O&M total roads operations and maintenance in 2014-16 was spending. Finally, the Phnom Penh City Hall executes funded by the national budget and was implemented by 23.7 percent of total public spending in maintenance, not MPWT (in charge of national and provincial roads, 33.6 far from the budget the rest of sub-national authorities percent of total) and by MRD (in charge of rehabilitation count with to maintain all rural roads in the country. The and emergency maintenance of rural roads, 11 percent estimate cost per kilometer of road construction works of total). Appointed local governments at the province and maintenance, by surface type, is presented in Table and district level receive 27.1 percent of total spending 40 in the Annex. Figure 70. About 87 percent of the national and provincial roads were in good or fair condition in 2017 Road condition for national and provincial roads, 2017 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 1-Digit NR 2-Digit NR PR (3-& 4-Digit) Good Fair Poor Very Poor Source: RDCMU, MPWT. Figure 71. Maintenance funding and implementation is divided among different public agencies Who funds O&M? (Average 2014-16) Who does/lead O&M? (Average 2014-16) Local Govt: Province & Districted Local Govt: Sub-national budget, City/Urban budget, (Appointed), 27.1% Commune & Districted 29.4% 23.7% (Appointed), 2.3% Non-Transport, 11.0% Donor, 0.0% MRD, 11.0% City Hall, 23.7% National budget, Joint MPWT/MR, 46.9% MPWT, 33.6% 0.0% Source: National Treasury database 2011-16, CDC DP database 2011-16, and sub-national budget 2011-16. 136 Cambodia: A Public Expenditure Review The share of paved national and provincial roads Figure 72. National and provincial roads has significantly expanded over the past two years, have a share of about a third of the total while it is still below targets. In 2016, 34.9 percent maintenance spending in Cambodia of total maintenance funding was directed to national Share of O&M funding, 2016 and provincial roads, 23.7 percent to urban roads, and the rest to rural roads (Figure 72). If urban roads Urban Road, 23.7% are excluded, national and provincial roads, which Other Road, 1.8% represent 80 percent of paved roads in Cambodia (and Rural Road, 26.5 percent of total roads) are receiving 46 percent 12.7% of the maintenance funding, whereas rural roads, which represent 73.5 percent of total roads but just RR: T3 & T4, 2.5% National Road, 20 percent of paved roads, receive about 53 percent RR: T2, 4.9% 34.9% of the maintenance funding (Table 36). The per- population road maintenance spending on rural roads RR: T1, 19.4% (approximately US$6/person) is higher than that of national and provincial roads (about US$4.5/person). Source: National Treasury database 2011-16, CDC DP database This is because of larger rural road network and a lower 2011-16, and sub-national budget 2011-16. population density in rural areas. It is worth noting that, while the percentage of 2-digit national roads that is paved has significantly increased, from 38 percent in abovementioned increase in maintenance spending. 2015 to 70 percent in 2017, it remains below the 90 In fact, in 2016, spending in both rehabilitation and percent target set in the NSDP. periodic and routine maintenance was above the level that would be required to maintain the current The recent emphasis on rehabilitation and O&M, share of national and provincial roads in good and fair beyond requirements, has led to improving road conditions (Figure 73). While this has helped improving conditions. Based on surveys, national road condition the condition in recent years, in a context of limited has improved recently, from 15 percent of total that resources and an underdeveloped rural network, were in good condition in 2011 to 17 percent in 2014 spending over the optimal level could potentially have and to 36 percent in 2017. This has been driven by the higher social returns in rural roads. Table 36. Shares of paved road (percent) % of roads that are 2009 2012 2015 2017 Share in total paved Share of paved roads in 2017 total roads National Road 1-Digit 99.1% 93.7% 100.0% 100.0% 21.2% 3.7% National Road 2-Digit 30.2% 55.9% 38.0% 70.4% 33.2% 8.2% Provincial Roads 1.7% 15.1% 10.0% 30.4% 25.9% 14.7% (3- and 4-Digits) Rural Road 0.3% N.A. N.A. 4.7% 20.0% 73.5% Source: Overview on Transport Infrastructure Sectors in the Kingdom of Cambodia: 2009, 2012, 2015; MPWT; MRD. Cambodia: A Public Expenditure Review 137 Figure 73. The levels of spending in O&M of 2016 would help preserving current road conditions Share of national roads in good & fair condition (%) Rehabilitation, periodic & maintenance spending comparison, US$ million per year 100% 140 120 80% 100 60% 80 60 40% 40 20% 20 0% 0 Current Year 5 Year 10 Year 20 Rehabilitation Periodic & Routine Maintenance Do Nothing Optimal RONET´s optimal estimates 2016 spending Source: World Bank staff calculations based in RONET. Policy options efforts would need to be connected. The current lack of a sector strategy results in fragmentation of efforts and uncoordinated construction works. Preparing As mentioned above, the lack of a guiding sector a sector strategy or transportation masterplan in a framework, coupled with institutional fragmentation consultative manner and involving all the sector actors and a piecemeal approach to investment planning, would be the only way to guarantee the achievement implementation, and maintenance, results in sub- of the Rectangular Strategy objective of developing optimal road spending in Cambodia. The lack of a multimodal transportation. The strategy would draw sector masterplan, coupled with independent processes from the recently finalized logistics masterplan as a for project selection and maintenance at MPWT and departure point. The strategy would be also helpful MRD, as well as limited information sharing by these two in determining coordinated investment priorities, as it road agencies, result in uncoordinated network building would consider during its elaboration the potential for and maintenance efforts. This gets complicated by lack economic and social benefits. Thus, it can also help of integration of government and DP-funded capital improving the efficiency of the project selection and expenditure, and limited coordination among financiers. prioritization process across institutions. This section discusses a series of policy options to address these challenges and improve the value-for- In addition, improve knowledge sharing and foster money of public road spending in Cambodia. joint activities to mitigate institutional fragmentation. Limited coordination and communication among First, develop a transportation masterplan as a MPWT, MRD and subnational entities in charge of guiding sector framework that can help align the transportation functions is likely to result in inefficient efforts of the different actors. These different actors operations and maintenance. Promoting more joint (MPWT, MRD, sub-national entities, Phnom Penh City MPWT-MRD projects, joint capacity building workshops, Hall, Phnom Penh Port Authority, airport concessionaires, and database sharing (including on the status of the etc.) may have different primary mandate, but their roads), can help improve coordination and attain more success often relies on good coordination due to the effective results. integrated nature of transportation network – their 138 Cambodia: A Public Expenditure Review In order to minimize the impact of budget budget cycle restriction for government-funded fragmentation, the preparation of budget strategic projects would need to be lifted. As previously plans by different public agencies would need to discussed, the current budget execution timeline be coordinated and aligned with the transportation implies that every year the technical specifications of all masterplan. The definition of programs, subprograms, the projects need to be prepared in January (at the line activities, and indicators in the BSPs at MPWT, ministries) and be submitted at the same time for review MRD, PPCH, and subnational entities, would need by MEF. If some project spans three years, it is divided to be defined along the guidelines provided in the in three proposals and, thus, paperwork is submitted transportation masterplan. Ideally, MPWT and MRD three times. This results in excessive bureaucracy and would have a similar structure of subprograms delays, with activities often starting just a month ahead (construction, reconstruction, maintenance, training, of the rainy season and then having to be frequently etc.) and pursue measurable outputs. Improving the interrupted. Delays in project preparation, tendering, classification and registration of spending (construction, fund release, or implementation can easily inflate project reconstruction, rehabilitation, maintenance, etc.) would costs, impacted by the rains. For a more efficient use also be necessary to identify technical and funding of public funding, further flexibility in the budget cycle gaps, as well as to ascertain the economic returns of would need to be granted. projects. Finally, DP-funded projects would need to be included in the BSPs, at least for information purposes, Also, demanding specification books before the as they contribute to the achievement of the objectives project enters the annual budget cycle could help defined. This would also help improving coordination avoiding some of the existing delays that lead among DPs. to increased implementation costs. Currently, the project specification books are prepared during the first Second, revise the long-list of capital projects by two months of the budget year, which results in a delay applying public investment management filtering in the effective request for funding and beginning of criteria, to keep only those that are realistic and in implementation. This could be addressed by resting line line with national priorities. As mentioned above, the ministries to start preparing these books as soon as the large difference in terms of funding between the long-list shortlist of selected project has been approved by the of projects presented by line ministries and the actual Council of Ministers (usually around November). allocation (short list) may be the result of unrealistic planning or, on the other hand, it could instead reveal In terms of budget allocation, the recent emphasis underspending, under-maintenance and a formidable on operations and maintenance seems to have led funding gap for roads. Identifying the roots of this to a significant improvement in road conditions large divergence merits a deep assessment of the on- over the past three years and should continue to going portfolio, a revision of the project selection and be the focus. At the same time, the decline in DP- definition, and the inception of an integrated strategy for funded construction and rehabilitation projects transport sectors that factors in local and modal targets may need to be partly compensated by higher and objectives. Authorities could consider imposing the government spending. Any allocation of additional requirement of preparing at least a pre-feasibility study funding into construction or reconstruction of national for projects, as well as depurating the existing long-list. and provincial roads would need to take into the Consider also integrating projects submitted to the PIP account the underinvestment by DPs in recent years in in the long list, even if they are expected to count with recent years in Northeastern provinces. Should this not DP funding. be compensated by government funded investment in those provinces, remoteness and difficulties in access Third, for the government to be successful in to both public services and markets would perpetuate replacing DPs in road provision, the one-year inequality. Cambodia: A Public Expenditure Review 139 In addition, it would be important to build monitoring undertaken to discern whether current spending and evaluation (M&E) mechanisms to ensure the patterns are adequate. Under-maintenance of rural effectiveness of road works and maintenance roads seems to be a concern in practice but, at this efforts. It is worth noting that M&E at the subnational point, it is difficult to determine the size of the problem, level does not need to comprise all the complete set of due to the fragmentation of O&M efforts across actors, features required at the national level. For example, Lao the limited monitoring of results, and to the lack of rural PDR has developed the Provincial Road Maintenance road surveys. In addition, requests for rehabilitation or Management System (PRoMMS), which is a simplified O&M funding under MRD (in the long-list) seem to lack road asset management system compared to the filtering; therefore, whether there is a massive under- existing national-level Road Management System. maintenance respect to the list submitted to MEF, PRoMMS collects data on road characteristics, or the funding mismatch obeys to poor planning and location, surface, drainage, and structure conditions, as pre-selection of projects at MRD is something hard to well as socioeconomic data of population within 2km assess. Conducting a comprehensive survey of rural of the roads. PRoMMS data is used to support road road quality and usage would be the most practical way maintenance prioritization decision making. to address this issue. An alternative to the survey could be the adoption of new technologies, such as the use of Finally, a systematic and up-to-date survey of the Mapillary on motorbikes to map the rural road network quality and usage of rural roads would need to be and road conditions. Table 37. Policy options towards more efficient spending on roads Challenge Short-term policy options Medium-term policy options (1-2 years) (3+ years) Transportation infrastructure Develop a transportation sector strategy or provision remains fragment- masterplan to align the efforts of different ed across actors institutions Public investment planning Require a pre-feasibility study for projects in Integrate projects susceptible to remains is uncoordinated the “long-list” of candidates for government receive DP funding in the “long- and unrealistic funding list” Delays in funding and imple- Lift the one-year budget restriction for mentation result in increased capital investment, and require specification cost of undertaking works books before January during the rainy season Lack on information on rural Conduct a survey of the quality and usage Build simplified monitoring and roads condition prevent of rural roads or, alternatively, use new evaluation (M&E) mechanisms to from assessing adequacy of technologies for the purpose ensure the effectiveness of road funding works and maintenance efforts in rural areas Source: World Bank staff elaboration. 140 Cambodia: A Public Expenditure Review Bibliography Abbas, A. S., & Klemm, A. (2013). A Partial Race to the Contreras, D., & Rau, T. (2012). Tournament incentives Bottom: Corporate Tax Developments in Emerging for teachers: evidence from a scaled-up intervention and Developing Economies. IMF Working Paper in Chile. Econ. Dev. Cult, 219-246. WP/12/28. Cuadra, G., Sanchez, J., & Sapriza, H. (2010). Fiscal ADB. (2012). EFAP - Final Report and Assessment of Policy and Default Risk in Emerging Markets, Review School Scholarship Program. Phnom Penh. of Economic Dynamics. Elsevier for the Society for Economic Dynamics, vol. 13(2), 452-469. ADB. (2015). Cambodia Addressing the Skills Gap Employment Diagnostic Study. Dabla-Norris, E., Brumby, J., Kyobe, A., Mills, Z., & Papageorgiou, C. (2011). Investing in Public Alesina, A., Campante, F., & Tabellini, G. (2008). Why Investment: An Index of Public Investment Efficiency. is Fiscal Policy Often Procyclical? Journal of the IMF Working Paper. European Economic Association, MIT Press, vol. 6(5), 1006-1036. Duflo, E., Hanna, R., & Ryan, S. (2012). Incentives work: getting teachers to come to school. Am Econ Rev, Alonso, J. D., & Sánchez, A. (2011). Reforming 1241-1278. education fi nance in transition countries : six case studies in per capita financing systems. Washington, Egis, E. (2012). Training Needs Assessment Report: DC: The World Bank. Ministry of Water Resources and Meteorology. Phnom Penh: Asian Development Bank. Bargawi, O. (2005). Cambodia's Garment Industry - Origins and Future Prospects. No 13, Working Egis, E. (2014). Economics of Operation and Papers. Economics and Statistics Analysis Unit Maintenance of Irrigation Systems. Phnom Penh: (ESAU), Overseas Development Institute. Asian Development Bank. Benveniste, L., Marshall, J., & Araujo, M. C. (2008). Esfahani, H., & and Ramirez, H. (2003). Institutions, Teaching in Cambodia. Washington, DC.: World Infrastructure, and Economic Growth. Journal of Bank. Development Economics, 443-477. Cavalcanti, C. B., Gustavo, A. M., & Le, T. M. (2014). Evans, D. K., & Popova, A. (2016). What Really Works Measuring the Impact of Debt-Financed Public to Improve Learning. The World Bank Research Investment. World Bank Policy Research WP 6766. Observer, 242-270. World Bank. Fan, S., Gulati, A., & Thorat, S. (2008). Investment, CDRI. (2013). Cambodia's Development Dynamics: subsidies and pro-poor growth in rural India. Past Performance and Emerging Priorities. Phnom Agricultural Economics 39, 163-70. Penh, September 2013.: Cambodia Development Resource Institute. S. Watkins (Ed.). Frankel, J. A., & Razin, A. (1987). The Mundell-Fleming Model A Quarter Century Later: A Unified Exposition. Cerdan-Infantes, P., & Filmer, D. (2015). Information, International Monetary Fund Staff Papers, Vol. 34, Knowledge and Behavior: Evaluating Alternative No. 4, 567-620. Methods of Delivering School Information to Parents. Washington: World Bank. Frankel, J. A., Vegh, C. A., & Vuletin, G. (2013). On graduation from fiscal procyclicality. Journal of Chau, N., & Kanbur, R. (2006). The Race to the Bottom, Development Economics, Elsevier, vol. 100(1), 32- from the Bottom. Economica, New Series, Vol. 73, 47. No. 290, 193-228. Cambodia: A Public Expenditure Review 141 Ganimian, A., & Murnane, R. (2016). Improving Ilzetzki, E., Mendoza, E. G., & Végh, C. (2010). How Education in Developing Countries: Lessons From Big (Small?) are Fiscal Multipliers? NBER Working Rigorous Impact Evaluations. Review of Educational Papers 16479. National Bureau of Economic Research, 719-755. Research, Inc. Gertler, P., Patrinos, H. A., & Rubio, M. (2012). IMF. (2007). Reforming Budget Systems Laws. IMF Empowering parents to improve education: Fiscal Affairs Department. evidence from rural Mexico. Journal of Development Economics, 68-79. James, S. (2014). Effectiveness of Tax and Non-Tax Incentives and Investments: Evidence and Policy Glewwe, P., & Muralidharan, K. (2016). Improving Implications. Available at SSRN: https://ssrn.com/ Education Outcomes in Developing Countries: abstract=2401905. Evidence, Knowledge Gaps, and Policy Implications. Handbook of the Economics of Education, Volume James, S., & Van Parys, S. (2009). Why Lower Tax 5, 653-739. Rates May be Ineffective to Encourage Investment: The Role of The Investment Climate. University of Goñi, E., & Maloney, W. (2017). Why don’t poor countries Ghent Working Paper Series 2010/676. do R&D? Varying rates of factor returns across the development process.”. European Economic Jayne, T., & Rashid, S. (2013). Input subsidy programs in Review, 94, issue C, 126-147. sub-Saharan Africa: a synthesis of recent evidence,. Agricultural Economics, 547-562. Gonzalez, O. C., Koehler-Geib, F., Aviles, A. M., Coppola, A., Cunha, B., Vostroknutova, E., . . . Jayne, T., Mather, D., Mason, N., & Ricker-Gilbert, J. Pertillar, T. L. (2012). Public Expenditure Review for (2013). How do fertilizer subsidy programs affect Peru: Spending for results. Washington D.C.: World total fertilizer use in sub-Saharan Africa? Crowding Bank. out, diversion and benefit/cost assessments. Agricultural Economics 44, 687-703. Government of Ireland, Department of Finance. (2005). Guidelines for the Appraisal and Management of Kaminsky, G., Reinhart, C., & Vegh, C. A. (2004). When Capital Expenditure Proposals in the Public Sector. it Rains, it Pours: Procyclical Capital Flows and Macroeconomic Policies. NBER Working Papers Grubert, H., & Mutti, J. (2004). Empirical Asymmetries in 10780. National Bureau of Economic Research, Inc. Foreign Direct Investment and Taxation. Journal of International Economics 62, 337–58. Kim, J.-H., Darcy, M., & Biletska, N. (2015). Public Investment Management for Public Private Haque, M., & Kneller, R. (2008). Public Investment and Partnership: Analytical Framework and Assessment Growth: The Role of Corruption. Manchester: Centre Tool. Washington: World Bank. for Growth and Business Cycle Research Discussion Paper Series No. 98. . Kim, J.-H., Darcy, M., & Biletska, N. (2015). Public Investment Management for Public Private Hausmann, R., Hwang, J., & Rodrik, D. (2006). What Partnership: Analytical Framework and Assessment you export matters. Working Paper 11905. Tool. World Bank. Hecock, R. D., & Jepsen, E. M. (2013). Should Countries Klemm, A., & Van Parys, S. (2012). Empirical Evidence Engage in a Race to the Bottom? The Effect of Social on the Effect of Tax Incentives. International. Tax and Spending on FDI. World Development, Elsevier, vol. Public Finance 19 (3), 393–423. 44(C), 156-164. Le, T. M., Raballand, G., & Palale, P. (2013). Zambia: Hill, H., & Menon, J. (2014). Cambodia: Rapid Growth Reforming a Broken PIM System. World Bank in an Open, Post-conflict Economy. World Economy, country study. 37, 1649–1668. Li, Y., Zhang, W., Ma, L., Huang, G., Oenema, O., Zhang, IFPRI. (2012). The Impacts of Public Investment in and F., & Dou, Z. (2013). An Analysis of China’s Fertilizer for Agriculture: Synthesis of the Existing Evidence. Policies: Impacts on the Industry, Food Security and Washington D.C: International Food and Policy the Environment. Journal of Environmental Quality. Research Institute. 142 Cambodia: A Public Expenditure Review Lopez, R. (2005). Why Governments Should Stop Non- RGC. (2015). Public Financial Management Reform Social Subsidies: Measuring the Consequences Program Evaluation in Cambodia. Based on for Rural Latin America. University of Maryland at Public Expenditure and Financial Accountability College Park. Methodology. Phnom Penh: Royal Government of Cambodia. Ly, S., & Aldaz-Carroll, E. (2014). Cambodia economic update: Clear skies. A Cambodia Country Economic RGC. (2015b). Market Orientation and Enabling Update. Phnom Penh: World Bank. Environment for Industrial Development Cambodia (Industrial Development Policy 2015 – 2025). Phnom Marshall, J., Chinna, U., & Hok, U. (2012). Educ Asse Penh: Royal Government of Cambodia. Eval Acc. RGC. (2015d). Agricultural Extension Policy in Muralidharan, K. (2012). Long-Term Effects of Teacher Cambodia. Phnom Penh: Ministry of Agriculture Performance Pay: Experimental Evidence from Forestry and Food Security. India. JPAL Working Paper. Riera-Crichton, D., Vegh, C. A., & Vuletin, G. (2014). Nguyen, T. (2009). Information, Role Models and Procyclical and Countercyclical Fiscal Multipliers: Perceived Returns to Education: Experimental Evidence from OECD Countries. NBER Working Evidence from Madagascar. Cambridge, MA: Papers 20533. National Bureau of Economic Massachusetts Institute of Technology. Research, Inc. Rajaram, A., Le, M. T., Kaiser, K., Kim, J.-H., & Sohnesen, T. P. (2016). School inputs and student Frank, J. (2014). The Power of Public Investment outcomes in 3 and 8th grade in Cambodia. Memo. Management: Transforming Resources into Assets for Growth. Washington: World Bank. Tandon, P., & Fukao, T. (2015). Educating the Next Generation: Improving Teacher Quality in Cambodia. Rajaram, A., Le, T. M., Biletska, N., & Brumby, J. (2010). Directions in Development. Washington, DC: World A Diagnostic Framework for Assessing Public Bank. Investment Management. Washington: World Bank Working Paper 5397. Timmer, C. P., & McCullough, E. (2010). Fertilizer Subsidies: A Policy Perspective. Rashid, S., Dorosh, P., Malek, M., & Lemma, S. (2013). Modern input promotion in sub-Saharan Africa: Ung, L., Oung, B., Tep, P., Kann, P., & Pring, M. insights from Asian green revolution. Agricultural (2016). Draft study: School Financing and Planning Economics 44, 705-721. Mechanism for Good Governance. Phnom Pehn. RGC . (2015c). PFM Reform Program: PFM Evaluation Wiedemann, V., & Finke, K. (2015). Taxing investments in Cambodia. Ministry of Economy and Finance. in the Asia-Pacific region: The importance of cross- border taxation and tax incentives. ZEW Discussion RGC. (2008). Public Finance System Law. Papers 15-014, , ZEW -Zentrum für Europäische Wirtschaftsforschung / Center for European RGC. (2014). National Strategic Development Plan Economic Research. 2014-2018. World Bank. (2015). The Philippines Open Roads RGC. (2014-2016). Ministerial Letters to the Prime Platform: Promoting Roads to Prosperity with on- Minister copied to three ministries (Ministry of Public line information and feedback. Works and Transport, Ministry of Rural Development, Ministry of Water Resources and Meteorology), World Bank. (2006). Fiscal Policy for Growth and during 2014-16. Development: An Interim Report. RGC. (2015). Alignment of MOEYS Budget and Annual World Bank. (2008). New Approaches to Input Operation Plan with MoEYS’ Priority Reforms and Subsidies, extracted from the World Development Education Strategy Plan 2014-2018. Report, Agriculture for Development. Washington: World Bank. RGC. (2015). Public Expenditure and Fiscal Accountability Assessment. Cambodia: A Public Expenditure Review 143 World Bank. (2011). Cambodia - More efficient government spending for strong and inclusive growth: integrated fiduciary assessment and public expenditure review (IFPAPER). Washington, DC: World Bank. World Bank. (2012). Using Impact Evaluations to learn how to promote schooling and learning. Phnom Penh. World Bank. (2014). Republic of India: Accelerating Agricultural Productivity Growth. Washington D.C.: World Bank. World Bank. (2015). Investment incentives in Cambodia: quantitative analysis. mimeo. The World Bank Group. World Bank. (2015). Investment incentives in Cambodia: inventory and qualitative analysis. World Bank, Mimeo. World Bank. (2015). Maintaining High Growth: Cambodia Economic Update. Phnom Penh: World Bank. World Bank. (2015). Rethinking the use of tax incentives in East Asia and Pacific, in East Asia and Pacific Economy Update. The World Bank Group. World Bank. (2015). Romania Advisory Services on Strengthening Public Investment Management. Output 12 Public Investment Management Manual, Reimbursement Advisory Service (RAS) Report. World Bank. (2016). Cambodia: Program Budgeting Review. World Bank. (2016). Reaping Richer Returns: Public Spending Priorities for African Agriculture Productivity Growth. Washington, D.C. World Bank. (2016). World Development Report 2016: Digital Dividends. Washington D.C.: World Bank. World Bank. (2017). Cambodia: Sustaining Strong Growth for the Benefit of All. Systematic Country Diagnostic. Washington D.C.: World Bank. World Bank. (2018). Cambodia Education Sector: Public Expenditure Tracking and Quality of Service Delivery Survey. World Bank. (forthcoming). Cambodia: Program Budgeting Review. 144 Cambodia: A Public Expenditure Review Annex Additional tables Table 38. DP Investment Project Funding, by sector mill USD % of Total DP by sector % of GDP Category 2013 2014 2015 2013 2014 2015 2013 2014 2015 General Administra- 29.61 5.93 6.05 25.8% 6.8% 7.9% 0.19% 0.04% 0.03% tion and Judiciary Defense - - - - - - - - - Security (Interior) - - - - - - - - - Economic Services 678.41 565.98 522.76 89.2% 80.2% 80.1% 4.42% 3.42% 2.88% Agriculture 144.98 148.65 132.52 81.5% 70.6% 81.7% 0.94% 0.90% 0.73% Transport 383.94 310.18 263.80 101.4% 100.3% 91.8% 2.50% 1.87% 1.46% Other Economic 149.49 107.16 126.44 83.9% 64.8% 73.1% 0.97% 0.65% 0.70% Services Environment, incl 12.77 8.84 18.90 48.7% 43.8% 62.2% 0.08% 0.05% 0.10% climate change Rural Development 38.27 57.21 17.02 67.9% 78.6% 58.9% 0.25% 0.35% 0.09% Social Services 74.17 88.71 136.89 25.5% 28.3% 43.3% 0.48% 0.54% 0.76% Health 17.39 29.11 35.17 10.8% 18.5% 24.8% 0.11% 0.18% 0.19% Education 35.98 38.91 65.56 39.8% 36.1% 57.4% 0.23% 0.24% 0.36% Other Social 20.80 20.69 36.15 52.1% 43.4% 60.6% 0.14% 0.13% 0.20% Services Other 0.00 0.01 0.88 0.0% 0.1% 3.2% 0.00% 0.00% 0.00% Emergency & Food 2.24 14.00 5.29 11.8% 57.4% 36.8% 0.01% 0.08% 0.03% Aid Budget & BOP - - - - - - - - - support Total Calculated 835.48 740.68 707.79 66.7% 61.1% 63.4% 5.44% 4.48% 3.90% Source: Calculations by World Bank staff based on CDC database. Cambodia: A Public Expenditure Review 145 Table 39. DP Technical Assistance, by sector mill USD % of Total DP by sector % of GDP Functional 2013 2014 2015 2013 2014 2015 2013 2014 2015 Category General Administra- 89.9 76.0 76.2 78.3% 87.2% 100.1% 0.59% 0.46% 0.42% tion and Judiciary Defense - - - - - - - - - Security (Interior) - - - - - - - - - Economic Services 66.3 83.1 99.8 8.7% 11.8% 15.3% 0.43% 0.50% 0.55% Agriculture 39.5 40.5 31.6 22.2% 19.2% 19.5% 0.26% 0.24% 0.17% Transport 8.2 7.4 25.5 2.2% 2.4% 8.9% 0.05% 0.04% 0.14% Other Economic 18.6 35.3 42.7 10.4% 21.3% 24.7% 0.12% 0.21% 0.24% Services Environment, incl 16.3 12.2 13.4 62.1% 60.2% 44.1% 0.11% 0.07% 0.07% climate change Rural Development 16.1 10.1 9.5 28.6% 13.9% 32.9% 0.11% 0.06% 0.05% Social Services 209.4 193.2 163.2 72.0% 61.7% 51.7% 1.36% 1.17% 0.90% Health 130.2 116.9 90.4 81.2% 74.1% 63.7% 0.85% 0.71% 0.50% Education 53.1 43.4 41.2 58.7% 40.3% 36.1% 0.35% 0.26% 0.23% Other Social Ser- 26.2 32.9 31.6 65.7% 68.9% 52.9% 0.17% 0.20% 0.17% vices Other 10.2 10.3 18.0 89.4% 104.7% 65.3% 0.07% 0.06% 0.10% Emergency & Food 1.4 1.4 0.1 7.4% 5.7% 0.3% 0.01% 0.01% 0.00% Aid Budget & BOP 0.0 13.3 8.1 - - - 0.00% 0.08% 0.04% support Total Calculated 409.7 399.5 388.3 32.7% 32.9% 34.8% 2.67% 2.41% 2.14% Source: Calculations by World Bank staff based on CDC database. 146 Cambodia: A Public Expenditure Review Table 40. Road construction and maintenance unit cost estimates Capital Road Works Unit Costs Two-Lane Unit Costs of Road Works ($/km) Surface Current Road Work Road Work Primary Secondary Provincial Type Condition Class Type Road Cement Good Condition Periodic Preventive Concrete Maintenance Treatment Fair Condition Resurfacing (Overlay) Poor Condition Rehabilitation Strengthening 850,000 850,000 680,000 (Overlay) Very Poor Con- Reconstruc- 800,000 800,000 640,000 dition tion No Road New Con- New Con- 850,000 850,000 680,000 struction struction Asphalt Mix Good Condition Periodic Preventive Maintenance Treatment Fair Condition Resurfacing 170,000 170,000 170,000 (Overlay) Poor Condition Rehabilitation Strengthening 190,000 190,000 152,000 (Overlay) Very Poor Con- Reconstruc- 250,000 250,000 200,000 dition tion No Road New Con- New Con- 300,000 300,000 240,000 struction struction Surface Good Condition Periodic Preventive Treatment Maintenance Treatment Fair Condition Resurfacing 35,000 35,000 35,000 (Reseal) Poor Condition Rehabilitation Strengthening 140,000 140,000 112,000 (Overlay) Very Poor Con- Reconstruc- 161,000 161,000 128,800 dition tion No Road New Con- New Con- 182,000 182,000 145,600 struction struction Gravel Good Condition Periodic Spot Regrav- Maintenance elling Fair Condition Regravelling Poor Condition Rehabilitation Partial Recon- 63,000 63,000 struction Very Poor Con- Full Recon- 74,000 74,000 dition struction No Road New Con- New Con- 93,000 93,000 struction struction Source: World Bank staff calculations using MPWT data. Cambodia: A Public Expenditure Review 147 Table 40. Continues Recurrent Maintenance Works Unit Costs Two-Lane Unit Costs of Road Works ($/ Surface Road C Road Work Road Work km-year) Type ondition Class Type Primary Secondary Provincial Road Asphalt Mix Very Good Recurrent Recurrent 1,100 1,100 1,100 Maintenance Maintenance Good Recurrent 1,200 1,200 1,200 Maintenance Fair Recurrent 2,500 2,500 2,500 Maintenance Poor Recurrent 5,000 5,000 5,000 Maintenance Very Poor Recurrent Maintenance Surface Very Good Recurrent Recurrent 2,500 2,500 2,500 Treatment Maintenance Maintenance Good Recurrent 2,700 2,700 2,700 Maintenance Fair Recurrent 3,000 3,000 3,000 Maintenance Poor Recurrent 5,000 5,000 5,000 Maintenance Very Poor Recurrent Maintenance Gravel Very Good Recurrent Recurrent 2,500 2,500 2,500 Maintenance Maintenance Good Recurrent 2,500 2,500 2,500 Maintenance Fair Recurrent 2,500 2,500 2,500 Maintenance Poor Recurrent Maintenance Very Poor Recurrent Maintenance Source: World Bank staff calculations using MPWT data. 148 Cambodia: A Public Expenditure Review Cambodia: A Public Expenditure Review 149 150 Cambodia: A Public Expenditure Review Cambodia: A Public Expenditure Review 151 152 Cambodia: A Public Expenditure Review Cambodia: A Public Expenditure Review 153 154 Cambodia: A Public Expenditure Review Cambodia: A Public Expenditure Review 155 156 Cambodia: A Public Expenditure Review Cambodia: A Public Expenditure Review 157 158 Cambodia: A Public Expenditure Review Cambodia Country Office Exchange Square Building Floor 10th IBRD and 11th IFC Streets 51-61 and streets 102-106 Sangkat Wat Phnom, Khan Daun Penh Phnom Penh, Cambodia www.worldbank.org/cambodia