Afghanistan in Transition: Looking Beyond 2014 Volume 2: Main Report May 2012 Contents Contents ............................................................................................................................................ i Preface .............................................................................................................................................. v Acknowledgments ............................................................................................................................ vi Acronyms ........................................................................................................................................ vii Chapter 1 Lessons from History—Afghanistan and Elsewhere .............................................................1 Summary ....................................................................................................................................................... 1 Some history ................................................................................................................................................. 1 International comparisons ............................................................................................................................ 5 Conclusions and recommendations ............................................................................................................ 19 References .................................................................................................................................................. 20 Chapter 2 The Economic Impacts of Transition .................................................................................. 21 Summary ..................................................................................................................................................... 21 Introduction ................................................................................................................................................ 21 The economy............................................................................................................................................... 22 Aid dependency .......................................................................................................................................... 27 Impact of transition on economic growth .................................................................................................. 30 Risks to macroeconomic stability................................................................................................................ 33 Impact of transition on employment .......................................................................................................... 36 Impact of transition on poverty .................................................................................................................. 39 Conclusions and recommendations ............................................................................................................ 42 References .................................................................................................................................................. 45 Chapter 3 Managing the Fiscal Challenge .......................................................................................... 46 Summary ..................................................................................................................................................... 46 Introduction ................................................................................................................................................ 46 Fiscal trends since 2001 .............................................................................................................................. 47 Medium-term projections........................................................................................................................... 52 Closing the financing gap ............................................................................................................................ 57 Conclusions and recommendations ............................................................................................................ 60 References .................................................................................................................................................. 62 Chapter 4 Building Government Capacity: Key Issues and Emerging Challenges ................................. 63 Summary ..................................................................................................................................................... 63 Introduction ................................................................................................................................................ 63 International experience............................................................................................................................. 65 Building a more effective Afghan state....................................................................................................... 67 Impact of transition .................................................................................................................................... 76 Conclusions and recommendations ............................................................................................................ 78 References .................................................................................................................................................. 80 Chapter 5 Delivering Services and Maintaining Infrastructure ........................................................... 82 Summary ..................................................................................................................................................... 82 Introduction ................................................................................................................................................ 82 Government presence and service delivery in three districts .................................................................... 84 Service delivery in five key sectors ............................................................................................................. 87 i Sustaining service delivery .......................................................................................................................... 99 Conclusions and recommendations .......................................................................................................... 105 References ................................................................................................................................................ 106 Chapter 6 Conclusions and Recommendations ................................................................................ 108 Main findings ............................................................................................................................................ 108 Appendix 1 MAMS for Afghanistan: Model Structure and Simulations ............................................. 112 Appendix 2 Fiscal Technical Background ......................................................................................... 132 Appendix 3 Range of civil servants’ remuneration mechanisms 2011 ............................................... 135 Figures Figure 1.1 Trends in per capita GDP growth, Afghanistan (2001 = year zero) and three clusters ............... 8 Figure 1.2 Extremely high and growing aid dependence.............................................................................. 9 Figure 1.3 Changes in life expectancy (top) and under-five mortality (bottom) ........................................ 13 Figure 1.4 Deteriorating rule of law (top) and government effectiveness (bottom) indicators................. 15 Figure 1.5 Deteriorating political stability indicator ................................................................................... 16 Figure 2.1 Real and agricultural GDP growth since 2003/04 (percent) ...................................................... 22 Figure 2.2 Aid trends, Afghanistan.............................................................................................................. 23 Figure 2.3 Sector growth rates, 2003/04–2010/11 .................................................................................... 24 Figure 2.4 Sector contribution to GDP growth (percent) ........................................................................... 25 Figure 2.5 Aid versus GDP per capita, average of 2006–08 ........................................................................ 27 Figure 2.6 Government consumption versus GDP per capita, average of 2006–08................................... 29 Figure 2.7 Decomposition of GDP at factor cost by activity and simulation, 2010/11–2018/19 (percent)32 Figure 2.8 Balance of payments, 2002/03–2010/11 (percentage of GDP) ................................................. 34 Figure 2.9 Money reserves, 2002/03–2010/11 .......................................................................................... 35 Figure 2.10 Unemployment and underemployment by province .............................................................. 37 Figure 2.11 Distribution of the employed population by economic sector, 2007/08 (percent) ................ 39 Figure 2.12 Provincial relationship between conflict and spending ........................................................... 39 Figure 2.13 Provincial spending and development outcomes, 2005 and 2007–08 .................................... 40 Figure 3.1 Aid trends, Afghanistan (US$ billion) ........................................................................................ 47 Figure 3.2 Breakdown of domestic revenues, 2010/11 (percent) .............................................................. 48 Figure 3.3 Breakdown of customs duties, 2010/11 (percent) .................................................................... 48 Figure 3.4 Public spending—external and core budget, 2010/11 (US$ billion) ......................................... 49 Figure 3.5 Core budget development and operating expenditure (percentage of GDP) ........................... 50 Figure 3.6 Core budget by sectors, 2011/12 (percent) ............................................................................... 51 Figure 3.7 Breakdown of revenue from large taxpayer offices, 2010/11 ................................................... 53 Figure 3.8 Projected expenditure by 2021/22 (percentage of GDP) .......................................................... 54 Figure 3.9 Projected expenditure by 2021/22 (2011/12 US$ million) ....................................................... 55 Figure 3.10 Core budget projections to 2021/22 (percentage of GDP) ...................................................... 56 Figure 3.11 Operating and total core financing gap (percentage of GDP) ................................................. 58 Figure 3.12 Total core financing gap (in 2011/12 US$ million) .................................................................. 59 Figure 3.13 Core operating and development budget execution, 2005/06–2010/11 (percent and USUS$ million) ........................................................................................................................................................ 60 Figure 4.1 Share of externally funded staff in eight ministries and one agency, 2011 (percent)............... 73 Figure 4.2 Share of EFS in total payroll costs in surveyed ministries and agency ...................................... 74 Figure 5.1 Core education budget by expenditure type, 1390 (2011/12) (US$ million) ........................... 88 ii Figure 5.2 Education development budget by donor, 1390 (2011/12) (US$ million) ............................... 89 Figure 5.3 Core health budget by expenditure type, 1390 (2011/12) (US$ million) ................................. 91 Figure 5.4 Health development budget by donor, 1390 (2011/12) (US$ million) ..................................... 91 Figure 5.5 Core rural livelihoods budget by expenditure type, 1390 (2011/12) (US$ million).................. 93 Figure 5.6 Rural livelihoods development budget by donor, 1390 (2011/12) (US$ million) ..................... 94 Figure 5.7 Core transport budget by expenditure type, 1390 (2011/12) (US$ million) ............................ 96 Figure 5.8 Transport development budget by donor, 1390 (2011/12) (US$ million) ................................ 96 Figure 5.9 Electricity core budget by expenditure type, 1390 (2011/12) (US$ million) ............................ 98 Figure 5.10 Electricity development budget by donor, 1390 (2011/12) (US$ million) .............................. 99 Figure 5.11 Core government budget, 1390 (2011/12) allocations and 1389 (2010/11) expenditure (US$ million) ...................................................................................................................................................... 100 Figure A1.1 Aggregated payment flows for MAMS, Afghanistan ............................................................. 112 Figure A1.2 Expenditure shares in the government and donor budgets, 2010/11 (percent) .................. 116 Figure A1.3 Assumed on-budget aid and external aid developments in the BASE scenario (gradual decline), 2009/10–2025/26 (2009/10 USUS$ billion) .............................................................................. 117 Figure A1.4 Assumed on-budget aid and external aid developments in the AID- scenario (rapid decline), 2009/10–2025/26 (2009/10 USUS$ billion) ............................................................................................. 123 Figure A1.5 Additional mining value added in BASE and MIN+ (constant US$ million) .......................... 125 Tables Table 1.1 Afghanistan’s performance relative to the strong performers cluster ......................................... 5 Table 2.1 Sector shares in GDP, percent ..................................................................................................... 24 Table 2.2 Donor assistance (US$ million, unless otherwise indicated)...................................................... 28 Table 2.3 Domestic economic content of aid, 2010/11 (percent) .............................................................. 30 Table 2.4 Real macro indicators by simulation, 2010/11–2018/19 (percentage annual growth) .............. 31 Table 2.5 Change in key outcomes relative to low-spending provinces, 2005–08..................................... 41 Table 2.6 Change in key outcomes after controlling for conflict, 2005–08 ................................................ 41 Table 2.7 Disaggregated impacts of high spending on food consumption groups, 2005–08 (Af per month) ............................................................................................................................................ 42 Table 4.1 Grade breakdown of pay and graded civil servants at the center and provinces ...................... 70 Table 4.2 Total donor assistance and technical assistance......................................................................... 71 Table 4.3 EFS in key ministries and agencies .............................................................................................. 73 Table 4.4 Salaries paid by donors for similar positions (US$ ) .................................................................... 74 Table 4.5 EFS employed by key programs and projects ............................................................................. 75 Table 5.1 National priority programs .......................................................................................................... 83 Table 5.2 Expenditures and costs, three districts, 2010–11 (US$ million, unless otherwise indicated) ... 86 Table 5.3 O&M needs for education infrastructure (US$ million) ........................................................... 101 Table 5.4 O&M needs for health infrastructure (US$ million)................................................................. 102 Table 5.5 O&M needs for rural livelihoods infrastructure (US$ million) ................................................. 103 Table 5.6 O&M needs for road infrastructure (US$ million).................................................................... 103 Table 5.7 O&M needs for electricity (US$ million) .................................................................................. 104 Table 5.8 Summary of transition challenges for five key sectors (US$ million)....................................... 104 Table A1.1 Descriptions of scenarios ........................................................................................................ 114 Table A1.2 Real macro indicators by simulation (percentage annual growth)......................................... 119 Table A1.3 Average annual growth in value added, by activity and simulation, 2011/12–2025/26 (percent) ................................................................................................................................................... 120 iii Table A1.4 Shares of GDP at factor cost by activity in first report year and by simulation in final report year (percent) ........................................................................................................................................... 120 Table A1.5 Decomposition of growth in GDP at factor cost by activity and simulation from first to final report year (percent) ................................................................................................................................ 121 Table A1.6 Domestic demand assumptions of different types of aid....................................................... 122 Table A1.7 ................................................................................................................................................. 127 Table A1.8 ................................................................................................................................................. 128 Table A1.9 ................................................................................................................................................. 128 Table A1.10 ............................................................................................................................................... 129 Table A1.11 ............................................................................................................................................... 130 Boxes Box 1.1 Somalia’s experience...................................................................................................................... 11 Box 2.1 The opium economy ...................................................................................................................... 26 Box 2.2 Mining as a potential source of growth ......................................................................................... 26 Box 2.3 Overview of scenarios and assumptions........................................................................................ 30 Box 2.4 Limitations of the model ................................................................................................................ 33 Box 4.1 Some key international lessons for Afghanistan............................................................................ 66 Box 4.2 Capacity Building for Results Facility.............................................................................................. 77 Box 5.1 Key findings on service delivery in three districts .......................................................................... 85 iv Preface It has been said many times that Afghanistan is at a crossroads. This has never been truer than now. The withdrawal of most international troops by 2014 will have a profound and lasting impact on the country’s economic and development fabric. This study explores some of these ramifications. This report provides the analytical and quantitative underpinnings to the executive summary and PowerPoint presentation on Transition disseminated by the World Bank in November, 2011 prior to the International Conference on Afghanistan in Bonn in December, 2011, and marks the culmination of the first of a set of studies examining different aspects of Transition in Afghanistan. Other studies under preparation for presentation at the Tokyo conference on Afghanistan in July 2012 will look in greater detail at what different levels of aid can buy in terms of development outcomes; From “Transition to Transformation —a more in-depth analysis of Afghanistan’s development financing needs through to 2025; and a Resource Corridors strategy, which aims at leveraging the forthcoming mining sector investments to ensure they benefit the broader economy. This first report is intended to be comprehensive, so it also discusses the broader historical and political economy context of development in the country, and how Afghanistan compares with other countries that have undergone their own transitions over the past 30 years. While many features of the Afghan story and the current challenges that Afghanistan faces are unique, other countries share different elements of that story—and may offer lessons on how to move away from violence and establish an enduring and stable transition to a better future. Finally, it is well known that collecting reliable data on Afghanistan is extremely difficult. Moreover, much of the information that is available is subject to large margins of uncertainty, as well as often problems of incompleteness, incomparability, etc. Data are frequently changed and updated. Collecting and triangulating data on issues such as jobs, aid inflows, and security costs has posed a major challenge for the team. This report is based on data collected from various sources in 2011, and its analysis and findings therefore comprise the team’s considered assessment using the best available information available by the end of that year. In addition, projections of future trends in Afghanistan inevitably are subject to uncertainty and reflect any weaknesses in the underlying data. Thus the report’s projections should be seen as subject to further adjustments and improvements as better and more recent information become available. For example, a lower targeted level for the Afghan National Security Forces (ANSF) is currently being discussed by NATO and may be announced at the NATO Security Conference in Chicago in May 2012, but this should not change one of the main conclusions of this report that they will still need to be mostly covered by international funding for many years. The subsequent reports for Tokyo will update some of the data based on more recently collected information in 2012, and some of the projections may be refined accordingly. Whenever we use data, we have been careful to specify its source. This report is presented in two volumes. Volume I is a stand-alone Overview which highlights the main findings, projections, and recommendations of the study. Volume II consists of five chapters presenting the detailed empirical background, analytical findings, projections, and recommendations of the study, along with a concluding chapter and three technical Appendices. v Acknowledgments The report forms part of a broader examination by the World Bank of transition in Afghanistan. The team wishes to thank the Government of Afghanistan for its support to this analytical work. Partial funding for this study was provided by the U.K. Department for International Development and Australian Agency for International Development, whose contributions are gratefully acknowledged. The report was edited by Communications Development. The study is the product of the combined efforts of a core team led by Richard Hogg (Governance adviser) and including Claudia Nassif (senior country economist), Camilo Gomez Osorio (economist), and Andrew Beath (economist) in Kabul, and William Byrd (consultant) in Washington, DC. Valuable contributions were made by: Dean Mitchell Jolliffe, Silvia Redaelli, Susanna Gable, Hans Lofgren, Richard Bontjer, Richard Scarth, Dallas Newby, Sakuntala Akmeemana, Kenneth Anye, Fotini Christia, Kinga Krisko, Gary Milante, Dhanie Nugroho, Khaled Payenda, Maha Ahmed ,Satyendra Prasad, Samina Bhatia, Gretchen Biery, Alice Burt and Astri Suhrke. The team benefited from comments by Brian Pinto, Aart Kraay, Shanta Devarajan, Hugh Riddell, Alexandre Marc, Graham Teskey and Nick Manning, who carefully reviewed the work at several stages during the preparation. The team has benefited from the support of the Afghanistan Country Management Unit in Kabul, Nicholas Krafft (former Country Director), Josephine Bassinette (acting Country Director), and the guidance provided by the South Asia Region PREM Management team, consisting of Ernesto May, Joel Hellman, Tony Verheijen and Vinaya Swaroop. The work was refined over many presentations to the government and donors in 2011. To this end we would like to thank the Minister of Finance, Dr. Omar Zakhilwal and the head of the Afghanistan Transition Office, Dr. Ashraf Ghani. A PowerPoint and short executive summary on the findings of the study were formally presented at the International Conference on Afghanistan in Bonn in December 2011. Authors of the individual chapters include William Byrd: chapter 1 “Lessons from History—Afghanistan and Elsewhere”; Claudia Nassif: chapter 2 “The Economic Impacts of Transition”; Camilo Gomez Osorio: chapter 3 “Managing the Fiscal Challenge”; Richard Hogg: chapter 4 “Building Government Capacity”; Andrew Beath and William Byrd: chapter 5 “Delivering Services and Maintaining Infrastructure.” Finally, the team would like to thank the international donor community in Kabul and ISAF as well as sector colleagues in the Kabul World Bank office for providing useful comments and valuable data at various stages. vi Acronyms ADB Asian Development Bank ANSF Afghan National Security Forces ARTF Afghanistan Reconstruction Trust Fund BPHS Basic Package of Health Services CBR Capacity building for results CRS Congressional Research Services DAC Development Assistance Committee (of the OECD) EFS Externally funded staff GDP Gross domestic product GNI Gross national income HIPC Heavily Indebted Poor Country IARCSC Independent Administrative Reform and Civil Service Commission ISAF International Security Assistance Force MAIL Ministry of Agriculture, Irrigation and Livestock MCP Management Capacity Program MEW Ministry of Energy and Water MoPH Ministry of Public Health MPW Ministry of Public Works MRRD Ministry of Rural Rehabilitation and Development NPP National priority program NSP National Solidarity Program O&M Operation and maintenance OECD Organization for Economic Co-operation and Development P&G Pay and grading PREM Poverty Reduction and Economic Management PRT Provincial reconstruction team UN United Nations UNDP United Nations Development Program USAID U.S. Agency for International Development WGI Worldwide Governance Indicators vii Chapter 1 Lessons from History—Afghanistan and Elsewhere Summary Afghanistan faces the transition with strengths and weaknesses after mixed progress over the past decade: rapid economic growth, low inflation, robust public financial management, some effective national programs, and improvements in social indicators (but from a very low base), set against low and deteriorating governance-related indicators and very high aid dependency. International experience and Afghanistan’s own history show that an abrupt cutoff in aid can lead to fiscal crisis, loss of control over the security sector, collapse of political authority, and possibly civil war (as in Somalia after 1988 and Afghanistan itself after 1991). They also suggest that political stability and state consolidation (based on building “inclusive enough” coalitions) are crucial for successful transitions (as seen in Cambodia, Mozambique, and Rwanda). Experience and history point to the critical importance of national leadership in successful transitions from conflict to peaceful, stable development. Such leadership outweighs the influence of international actors. Regional factors among neighboring countries can exacerbate and perpetuate conflicts (as in the Democratic Republic of Congo and Somalia); regional actors and those further afield need to maximize efforts to ensure that such factors do not undermine Afghanistan’s transition. Effective transitions are generally associated with robust economic growth (for example, Mozambique and Rwanda) and less successful transitions with slower or negative growth. Some history History—Afghanistan’s and that of other conflict-affected countries—holds important lessons for the transition. These lessons should not be mechanically applied, as every country’s context is unique, but they offer insights for policy makers. Afghanistan was born as a state in the mid-18th century, formed as an expansionist Pashtun-led dynastic power under founder Ahmad Shah Durrani (1747–72) that conquered Delhi and took over parts of the Indian subcontinent. It was molded into its present territorial boundaries during a century of wars and diplomacy involving geopolitical rivalry between British interests in India to the east and south, Russia expanding from the north, and to some extent Persia (now the Islamic Republic of Iran) to the west. There were three Anglo–Afghan wars, numerous violent incidents and incursions, and many internecine revolts and succession struggles. Afghanistan became a buffer state between the British and Russian empires. 1 Abdur Rahman Khan (1880–1901), who came to power at the end of the Second Anglo–Afghan War, forcibly asserted control over Afghanistan’s present territory through a series of “internal invasions” and started to build basic institutions of a central state. After several decades of semi-isolation after the Third Anglo–Afghan War in 1919—including aborted reform under King Amanullah (1919–29) and neutrality during World War II—Afghanistan’s geopolitical role became important again during the Cold War, when it was one of the world’s highest per capita aid recipients. The country’s geopolitical context was complicated by the creation of Pakistan in 1947, which came to see its interests in Afghanistan through the lens of its rivalry with India. The country modernized slowly over the next three decades, but remained very poor with extremely weak social indicators. After a long period of relative peace and stability under King Zahir Shah (1933– 73), it saw increasing instability: a bloodless coup led by the King’s cousin Mohammad Daoud in 1973, which instituted a republic; a bloody communist coup and takeover in 1978; and the Soviet occupation at the end of 1979, which led to two decades of protracted and debilitating conflict. Afghan national identity has been formed and defined by resistance against foreign incursions, which often took on a religious dimension because most incursions in recent history were by non-Muslim powers. Afghanistan has a longer history as a distinct entity than most of its neighbors—for example Pakistan, established in 1947, and the Central Asian states, which were formed after the collapse of the Soviet Union in 1992. Moreover, despite ethnic diversity and problems exacerbated by Pashtun dominance and interethnic conflict, Afghanistan has never had a serious separatist movement. During the long reign of Zahir Shah the country showed it could be politically stable. Yet the pre-1973 monarchical state did not penetrate deeply into the countryside. While in theory Afghanistan has always been a unitary state, local governance has been largely traditional and informal. At its best the system provided a workable compromise between centralized monarchical rule and the diverse, decentralized, and traditional reality in most of the country. The state upheld order, maintained reasonable control over borders, exercised independent diplomacy in a difficult region, and was perceived as legitimate internally and externally—all basic state functions that subsequent governments have struggled to fulfill. A key lesson from Afghanistan’s 20th century experience is that overly ambitious and rushed efforts to modernize have triggered sharp domestic reactions that set back development, sometimes for decades—whether the internal efforts of King Amanullah in the 1920s or the externally-driven measures of the Soviet-backed Communist regime. A gradual and evolutionary approach, in contrast (as in the 1950s–70s), achieved modest progress when it was initiated sensitively and did not directly threaten conservative elements in society. During most of its existence, Afghanistan has relied on external financing to run the state and make public investments. Yet large aid programs in the 1950s–70s did not spread benefits widely, let alone spur and sustain rapid economic development. In the 1970s it remained a very poor, largely subsistence- based agricultural economy with extremely low social indicators. But Afghanistan built infrastructure (most notably in the form of highways and some irrigation projects) and developed higher education— even if on a narrow, elite-based model. The opportunistic coup that brought a pro–Soviet Communist government to power in April 1978 was followed by not only ambitious reforms, but also the wholesale imprisonment, torture, and execution of opponents, leading to a widespread resistance that threatened the new regime and eventually led to the Soviet military intervention at the end of 1979. This marked an important juncture in Afghanistan’s history, and violent conflict, in various forms and to differing degrees, has continued to this day. During 1986–89 the Soviet Union looked for a way to exit militarily from the country. Key elements of the Soviet withdrawal and transition included: 2  Removing Babrak Karmal, seen as an ineffective head of state and Afghan Communist Party leader, in 1986, and replacing him with Najibullah.  Reversing the Communist reform agenda and switching to a National Reconciliation Policy, that called for a cease-fire with mujahideen forces, talks, a transitional government, and elections.  Further strengthening and enlarging Afghan security forces, which reached a peak of 400,000 after the Soviet military withdrew, and which showed considerable operational capability.  Seeking international agreement and cover for withdrawal through the United Nations, resulting in the Geneva Accords of April 1988; the United Nations was supposed to oversee implementation of the Geneva Accords, which were signed by Afghanistan and Pakistan and “guaranteed” by the Soviet Union and the United States—but the accords neither included, nor were signed by, mujahideen representatives.  The rapid and orderly withdrawal of Soviet military forces from May 1988 to February 1989, which was carried out without major problems. Mujahideen groups and their main external backers (including Pakistan, Saudi Arabia, and the United States) opposed the reconciliation policy, so the government was left to make deals with individual local leaders. The government provided financial incentives and reportedly achieved some success locally. The Soviet withdrawal occurred under difficult conditions, with little support from other parties and when the Soviet Union itself was weakening. Given these circumstances it can be considered a success in its own narrow terms, particularly in achieving the limited objectives of a face-saving exit and leaving behind a regime that did not immediately collapse. Many advisers, KGB personnel, and other Soviet officials remained in Afghanistan to support the government, and much military hardware was left behind for the Afghan Army. But the weaknesses of the Geneva Accords—which implicitly allowed the Soviet Union and the United States to continue providing arms and other help to both the government and the mujahideen—meant that the Soviet withdrawal was not supported by a meaningful peace settlement. After the withdrawal, Najibullah showed himself an astute politician by managing to cling to power for three years. His government held on to the large cities, contrary to widespread expectations at the time. The conflict took on more the characteristics of a civil war, which allowed Najibullah to exploit divisions within the mujahideen. The government co-opted a range of tribal leaders and negotiated deals for local control, which often took the form of government-backed militias paid with Soviet-provided aid including weapons, food, and fuel. This approach worked in the short term but carried serious risks, since those co-opted easily could (and did) change sides and turn against the government if funding was cut off—as happened in 1991 with the end of the Soviet Union—or if the opposing side offered a higher price. This cutoff in Soviet material and financial aid precipitated the regime’s collapse. The government could no longer pay its militias or army, and they either turned against the government or defected. Kabul fell in April 1992. The mujahideen regime that took over was unable to form a unity government, and the situation rapidly degenerated into a civil war that destroyed Kabul. This created the conditions for the Taliban movement to emerge and rapidly expand from 1994. It took over Kabul in 1996 and, by the end of the decade, controlled some 90 percent of the country. Despite its different background, context, and dynamics, the Soviet withdrawal and its aftermath may have some lessons for the current transition in Afghanistan:1 1 See Fotini (2011) and Byrd (2012). 3  Effective Afghan leadership was critical in the Soviet withdrawal and for the survival of the post- Soviet government, and is likely to be so in the current transition. This lesson is consistent with that from international experience on the importance of national leadership at critical points in transitions.  The withdrawal of foreign military forces removed a major justification for mujahideen opposition, allowing the government to exploit divisions among them.  Total reliance on Afghan security forces after foreign military withdrawal was unrealistic, especially without a meaningful peace agreement or at least a durable cease-fire. The post-Soviet Afghan army was well equipped and well officered (many officers had been trained in the Soviet Union), but was unable to do much more than hold on to the larger cities after the Soviet withdrawal.  The alternative of relying on local deals with militias and armed groups worked for a short time, but required continuing large injections of external funds and risked instability due to shifting alignments.  Durable peace was impossible without the buy-in of key regional players, notably Pakistan. (Pakistan was a signatory to the Geneva Accords, but the mujahideen groups based in Pakistan were not.)  The Geneva Accords did little more than provide a face-saving cover for the Soviet military withdrawal, and its shortcomings were evident with the failure to bring the mujahideen to the peace table.  The post-Soviet Afghan government depended heavily on Soviet financial and material support. Once this dried up the regime quickly collapsed. Future historians may see Afghanistan’s early 21st century history in four phases:  2001–05: optimism for the prospects of building a modern, developmental, and democratic state based on the Bonn Agreement of December 2001.  2005–09: Taliban resurgence and “militarization” of aid, which increasingly became a tool for short- term stabilization interventions in conflict-affected areas.  2009–11: U.S. military and civilian surge and a dramatic increase in the U.S. aid budget.  2011–14: transition, intended to mark the phased drawdown of international combat forces and the Afghan assumption of full sovereignty over security. Aid flows during 2001–05 were relatively modest compared with those to many other post-conflict countries, and the international military intervention was relatively small and limited in mandate. Facing an intensifying Taliban insurgency, international troops and aid increased after 2005 as many forces moved into the south of the country, with a further surge from 2008/09. Afghanistan now faces the planned drawdown of foreign combat forces and the handover of security to the Afghan National Security Forces (ANSF) by the end of 2014. Aid flows are expected decline from 2012 onward, and may fall faster after 2014. The security transition is scheduled to be completed by the time of the next Afghan election cycle, which includes presidential elections in 2014 and parliamentary elections in 2015. If the transition is accompanied by a reconciliation process, cessation of hostilities, and ultimately a peace agreement with the Taliban, it would turn out to be more akin to other, conventionally defined post-conflict transitions. Otherwise, it will largely involve an international drawdown of military forces and declining aid in the absence of any peace agreement. 4 International comparisons Afghanistan faces the next three years with both strengths and weaknesses after uneven progress in the past decade. Table 1.1 presents a broad summary of its performance since 2001 as measured by economic, social, and governance indicators (the latter perceptions based), against a group of strongly performing comparator countries—one of three “performance clusters” (weak, middle, and strong).2 It also summarizes the current gap with these strong performers, indicating how far Afghanistan needs to go to become something like a typically successful post-conflict, low-income country 15 years after the start of transition. These comparisons, and in particular the underlying data and indicators on which they are based, carry a significant margin of uncertainty and hence need to be interpreted cautiously, but nevertheless they are suggestive and provide insights on Afghanistan’s recent performance and current situation from an international perspective. Table 1.1 Afghanistan’s performance relative to the strong performers cluster Afghanistan Trend of strong Variable/indicator trend performers Current gap Comments Economic performance Growth in Afghanistan has been faster than in Real average per capita Very rapid Rapid None virtually all comparators; needs to maintain GDP growth robust growth even if at somewhat lower rates Average per capita real Avoiding recession and continuing growth, Rising sharply Rising None GDP level especially employment growth, will be key Needs to maintain good performance in Inflation (GDP deflator) Very low Not so low None controlling inflation during difficult transition Afghan structural change appears to a Structure of economy Rapid change Rapid change See comment considerable extent aid driven; needs new growth drivers Government Government consumption for Afghanistan Flat Flat Not significant consumption substantially higher due to external budget Human development Improving from It will take Afghanistan a very long time to catch Life expectancy Improving Large low base up with the strong performers’ average Limited Under-five mortality progress Improving Very large As previous comment rate starting from very low base Afghanistan is still behind, but trends are not Primary enrollment Improving from Improving Large closely associated with success of transition in ratio very low base comparators 2 These comparisons are based on an “event study” approach, whereby trends and levels for various indicators are compared across countries not for the same calendar years but rather over time following a specific transition initiation year (different for each country). In the case of Afghanistan, for the purpose of these comparisons the initial transition year selected is 2001, when the Taliban regime fell. The methodology for the international comparisons and clustering performance groups is discussed in more detail in Byrd (2012). 5 Afghanistan Trend of strong Variable/indicator trend performers Current gap Comments Governance and rule of law Reinforces widespread concerns about Rule of law indicator Deteriorating Improving Large worsening rule of law in Afghanistan; adverse trend needs to be reversed Government Declining after Declining trend is worrisome and needs to be Improving Large effectiveness indicator initial rise reversed; will be challenging with falling aid Corruption in Afghanistan is very serious, but Control of corruption Low and fairly Higher, variable there is much variation across countries Large indicator stable across countries (including strong performers) and no clear trends A recent decline for Afghanistan, but trends in Voice and Steady then Modest Large comparators are mixed during transition and accountability indicator deteriorating improvements not sharply upward Small positive Recent deterioration, and if it continues will Civil liberties and Improving, then Small but trend but levels place Afghanistan well below most comparators political rights indicator Deteriorating growing vary greatly Regulatory quality Improving Further improvements are needed, as it will Improving Large indicator somewhat help facilitate private sector growth Political dimension Afghanistan’s poor performance relative to Political stability Deteriorating Improving Very Large most comparators (in all clusters) underlines indicator the issue’s seriousness Aid dependency Aid (ODA per OECD Declining aid during transition may have serious Rising, now Falling steadily Extremely definition) as a impacts; an abrupt fall certainly would extremely high over time Large percentage of GNI Technical assistance Fluctuating and Technical assistance will decline; need to Fluctuating at (OECD definition) as >5 percent of Very Large improve its quality and the transition to core lower levels percentage of GNI GNI government capacity Source: World Bank staff assessments, based on World Development Indicators and Worldwide Governance Indicators (see Byrd, 2012). Note: Gray = no gap or small gap with comparators; light red = intermediate gap or mixed pattern; dark red = large or very large gap. The country has in some ways performed very well since 2001 relative to comparator countries in post- conflict transitions (even the strongly performing cluster), and the priority now is to protect the gains that have been made, which may be at risk during transition. The most notable are rapid economic growth, low inflation, solid and improving public financial management, and some effective national programs providing basic services. In other areas, such as primary education enrollment and life expectancy, Afghanistan still lags that of comparator countries because of the very low base in 2001, among the worst in the world. Although perceptions-based governance-related indicators, such as rule of law, political stability, voice and accountability, and corruption, need to be interpreted cautiously, Afghanistan’s performance has been weak and the gap with comparators is large and growing. Finally, aid dependency in Afghanistan has become uniquely high among comparator countries. Economic performance A striking feature of Afghanistan since 2001 is its very rapid economic growth, with real GDP rising at around 9 percent a year (with wide fluctuations). Per capita growth has been close to twice that of strongly performing comparator countries (figure 1.1) and higher than that of all individual comparator countries. The strong performers have generally had higher average per capita GDP growth than middle 6 performers, while most weak performers’ economies contracted. After increasing by more than 75 percent in real terms between 2002 and 2009, average per capita GDP climbed above the average of strong performers during their transition periods, even though that group includes some starting with much higher incomes and levels of development. Key questions about Afghanistan’s growth relate to its sources and sustainability. Even taking into account the stagnation of real GDP during the previous 25 years of conflict and the subsequent post- conflict economic rebound, growth—especially in the latter part of the past decade—appears to have been stimulated by enormous inflows of aid and international military spending. Afghanistan’s control over inflation has been commendable. Soon after 2001, the country ended the 1990s’ hyperinflation with a successful currency reform, and subsequently brought inflation down despite global food and energy price shocks and supply blockages in neighboring countries, which increased price volatility. There seems to be, however, little relationship between transition performance and inflation; many strong performers had high inflation in the early years of transition, and middle performers had the lowest inflation on average. Still, avoiding very high inflation, let alone hyperinflation—which Afghanistan has done since 2001—is a prerequisite for a sound economy that offers good growth and employment prospects. Large aid financing of Afghanistan’s budget deficits has been helpful in controlling inflation, as it avoids the need to use inflationary money creation for deficit financing. But as aid declines during the transition, budgetary pressures will increase, requiring a combination of difficult expenditure prioritization, strong expenditure controls, robust domestic revenue growth, and a higher share of aid on budget. 7 Figure 1.1 Trends in per capita GDP growth, Afghanistan (2001 = year zero) and three clusters Afghanistan Weak performers Middle performers Strong performers Growth trend Central African Republic Burkina Faso Cambodia Chad Burundi Lao PDR Congo, Dem. Rep. Ethiopia Liberia Côte d’Ivoire Guinea Mali Eritrea Nepal Mozambique Guinea-Bissau Niger Nicaragua Haiti Growth trend Rwanda Zimbabwe Tajikistan Growth trend Uganda Vietnam Growth trend 240 220 200 180 Average per capita GDP 160 140 120 100 80 60 40 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Years after start of transition Source: World Development Indicators; figures compiled in Byrd (2012). Note: Year zero is the base year, with GDP per capita indexed to 100 (for Afghanistan 2002 is the base year since data on a comparable basis are not available for 2001); average per capita GDP on vertical axis, years after start of transition on horizontal axis (2001 in the case of Afghanistan). Middle and especially strong performers among comparator countries saw substantial change in the broad structure of their economies during transition, as the share of agriculture in GDP declined and those of industry and services rose; poor performers, however, saw a stagnating industry, rising share of agriculture, and a declining share of services. 8 The broad direction of structural change in Afghanistan has been similar to that of strongly performing comparator countries, but the size of its shift to services has been more marked, with much less of a rise in industry’s share. This suggests that international financial inflows and the service activities they funded (such as transport and security services) have played a leading role in the structural changes that have occurred in the economy. Aid dependency Except in 2001, Afghanistan has been more aid dependent (measured as a ratio of aid to gross national income [GNI]) than the averages for all three comparator groups. Contrary to international experience— and worrisome for sustainability—aid dependency has soared since 2001. It is now around 50 percent (according to OECD data on official development assistance), which is extremely high relative to virtually all comparator countries (figure 1.2). Moreover, estimates by World Bank staff of total aid and international military expenditure contracted in Afghanistan suggest aid shares much higher than the OECD data, mainly because these estimates include assistance to the ANSF. Afghanistan has become an extreme outlier in its dependence on aid. Trends in aid dependence are downward for all three groups of comparator countries. Strong performers on average have initially higher aid dependency ratios, but steeper declines than the other two groups, partly reflecting more rapid economic growth. Figure 1.2 Extremely high and growing aid dependence 100% 90% 80% 70% Aid as a percentage ratio to GNI 60% 50% 40% 30% 20% 10% 0% 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Years after start of transition Source: World Development Indicators; figures compiled in Byrd (2012). Note: Aid as a percentage ratio to GNI on vertical axis; year after transition initiation on horizontal axis; gray diamond points are OECD-DAC data for Afghanistan, round black points are World Bank staff estimates. 9 Data on technical assistance (according to the OECD definition of “technical cooperation”) as a share of GNI for Afghanistan and comparator countries indicate below 5 percent of GNI for strongly performing comparators with only a couple of exceptions. This ratio tends to be even lower for middle and weak performers, and all three groups show declines over time. Afghanistan has seen wide fluctuations in this ratio, with an upward trend and levels above 5 percent in all but two years during the post-2001 period. Thus Afghanistan is more reliant on externally funded technical assistance than comparator countries, and becoming even more dependent (see chapter 4). And, given the higher aid estimates by World Bank staff (which cannot be broken down between technical assistance and other categories), actual funding for such assistance in Afghanistan must be much higher than the OECD data suggest. Dependence on external financing—aid or other financial inflows—is nothing new for Afghanistan: it received massive amounts of Soviet aid in the 1980s and early 1990s; it was one of the highest per capita aid recipients in the world during the 1960s and 1970s; Afghan rulers received subsidies from Britain during much of the 19th century; and in its earliest years plunder from the Indian subcontinent was a main source of financing for the Afghan state. Historically, the regime rarely had to mobilize large revenues from its own people to cover costs and provide services, so this aspect of the social contract between state and society has long been missing. Instead, the historical pattern was often to use external resources to “buy loyalty” and provide security and political stability. The key economic question for Afghanistan’s transition is whether large reductions in aid to other post- conflict or conflict-affected countries have adversely affected their economies, particularly in the short term. The evidence suggests that the impact of large aid reductions on economic growth has been less dramatic than might be expected. In Mozambique, a strongly performing low-income comparator country, aid was equivalent to 60– 80 percent of GNI during the first three years of transition, but declined sharply to less than 30 percent of GNI and remained in the 20–30 percent range during most of the remainder of the 15-year transition period. Yet real per capita GDP maintained average annual growth of over 4 percent. In Guinea-Bissau, a weakly performing comparator country, aid as a share of GNI fluctuated widely in the 23–79 percent range, peaking in year three of transition. Average annual per capita GDP growth was generally negative, but there is no discernible relationship between aid and GDP performance. In Eritrea, another weakly performing comparator, aid started low, was higher amid fluctuations during years 8–12 of transition (in the 30–54 percent of GNI range), then fell sharply to 9–12 percent of GNI in years 13–15. Again no discernible relation is evident between Eritrea’s large swings in aid and average annual per capita GDP growth, which was negative from year 9 onward but slightly less so during the final three years of the 15-year transition period. In Bosnia and Herzegovina, a middle-income post-conflict country, a marked but steady decline in aid started soon after the end of conflict. Aid fell from a peak of 57 percent of GNI in 1995 to 6–8 percent in 2002–04 and eventually to 2–3 percent in recent years. This was associated with a slowdown in average annual per capita GDP growth from well into double digits initially (reflecting the post-conflict economic rebound) to as low as 3 percent for a couple of years. However, growth picked up to 5–7 percent in 2004–08—a good performance for a middle-income country. After independence in Timor-Leste, aid was around 70 percent of GNI during 2000–03 and then declined steadily to 10 percent over the next five years. Real per capita GDP growth was negative in the initial years of falling aid but picked up later, reaching significant positive levels after the decline in aid was completed. These five examples suggest that the anticipated decline in aid to Afghanistan may not result in anything like economic collapse, though the following factors may have played a mitigating role in these countries. 10 First, in Mozambique and to some extent Bosnia and Herzegovina, aid fell early on in the post-conflict period, when other trends—in particular the typical post-conflict economic recovery—were still strong and could partly offset the decline in aid. Second, in Bosnia and Herzegovina the decline in aid, though large and fairly rapid, was steady and did not involve violent year-to-year fluctuations, while in Mozambique aid after the decline remained substantial as a share of GNI (through the rest of the period covered by the analysis). Third, in all these examples the peak levels of aid tended to be short-lived, so their economies did not become reliant on and adapt to such high aid inflows. Fourth, in Bosnia and Herzegovina, most initial aid supported one-off activities such as rebuilding infrastructure, returning refugees, and other short-term quick-impact efforts that did not need to be repeated. Fifth, in that country, other inflows (remittances and, to a lesser extent, foreign direct investment) cushioned the impact of declining aid. But Afghanistan’s economy may have become more adapted to high aid inflows. Also, the typical post- conflict economic boom that fueled growth in the initial post-2001 years has already waned (but a durable peace agreement could stimulate a second boom). Nevertheless, international experience suggests that the Afghan economy could absorb the aid reduction without suffering dramatic economic contraction— as long as the decline in aid is programmed and gradual, and not subject to sudden, violent fluctuations (see chapter 2). Other countries’ experience and Afghanistan’s own history demonstrate that abrupt cutoffs or sharp reductions in aid can be highly destabilizing from a fiscal, political, and security perspective, even if the direct economic effects are more modest. This is especially true of aid that funds large security forces, which can become ineffective or even turn against the government if funding ceases, especially if they are not paid. Somalia is a good example of a country that experienced an abrupt cutoff of aid in the late 1980s that led to fiscal crisis, desertion of security forces (who were no longer paid), and a rapid collapse of the state (box 1.1). The same was true of Afghanistan after the sudden cutoff of Soviet aid to the Najibullah regime after 1991, which resulted in widespread defections from pro-government militias and eventually the national army, leading to the government’s collapse. Box 1.1 Somalia’s experience Somalia has suffered greatly from violent conflict for more than two decades. The socialist government of Siyad Barre, which came into power after a coup in 1969, initially promoted a strong sense of nationalist unity. But clan divisions within Somali society created vulnerabilities, running alongside deep poverty, rapid population growth, and urban drift with a growing pool of unemployed young men. The Cold War gave the regime access to large volumes of foreign aid, allowing it to engage in massive patronage and to build a large army, which absorbed unemployed youth. But by 1980, after a disastrous war with Ethiopia in 1977–78, several weak clan-based insurgencies had started. During the 1980s Somalia became one of the most repressive and predatory regimes in Africa, and the government resorted to divide and rule tactics, further exacerbating clan divisions. The country fell into civil war in 1988, triggered partly by an agreement between Somalia and Ethiopia to stop providing sanctuaries for each other’s insurgents, which had the unintended effect of prompting the Somali insurgent movement to launch an all-out attack to establish a presence in northern Somalia. The government response was brutal, with systematic human rights abuses, precipitating a flow of 300,000–500,000 refugees to Ethiopia. Foreign aid was suspended in response (a move made easier by the waning of the Cold War). Since the Somali state and military forces depended on aid, this led to desertion of military units and a wholesale shrinkage and retreat of the state. Clan-based movements proliferated, with ready recruits from the ranks of army deserters. After the Barre government fell in 1991, these groups failed to form a national government and instead fought over the spoils. 11 The conflict in Somalia has gone through several stages since, including a disastrous U.S. military foray and withdrawal; a protracted UN peace operation that was far from fully successful (though it did stop the fighting in 1993–95); subsequent signs of declining violence and an emerging patt ern of “governance without government”; and major changes from 2006 that were accompanied by worsening violence, including the rise of a major Islamist movement. Somalia’s experience does have some positive aspects, most notably improved security and the emergence of reasonably effective informal governance in the sizable regions of Somaliland and Puntland. These reflected the lack of aid to generate aid dependency and, in particular, the ability of community leaders to prevent serious fighting and the rise of a war economy. But the successes achieved in these two areas remain fragile. Lessons from Somalia’s experience include the following:  “Divide and rule” tactics of a government striving to stay in power, as used by the Siyad Barre regime during its later years, can cause serious political and social damage, particularly in countries already characterized by strong clan, ethnic, or other divisions that are easy for power-holders to exploit.  Sanctuaries for insurgent groups can be a serious obstacle to successful counterinsurgency, but eliminating them (or threatening to do so), as was tried by Somalia in 1988, can have unintended adverse side effects such as pushing insurgents back into the country.  Abruptly stopping foreign aid, especially during a conflict, can expose the weakness of a highly aid-dependent state and lead to its collapse through loss of resources for patronage and associated political and security impacts (such as desertions from the army).  The emergence and entrenchment of a “war economy” can fuel the persistence of conflict, as various actors— even if originally motivated by grievances or political considerations—benefit from the spoils.  Overemphasis on building security forces and the relative neglect of civilian institutions, and a focus by the international community on numbers of security forces trained and equipped (rather than the genuine security of the country’s population), can be problematic.  Geopolitical significance (earlier in the Cold War or more recently in the U.S.-led “war on terror”) can complicate domestic groups’ efforts to negotiate peace. A focus only on power sharing in negotiations, without seeking genuine reconciliation or addressing underlying grievances, can be counterproductive. Source: Menkhaus 2011. Human development Human development indicators are an important measure of not only development but also a government’s ability to provide social services. For an overall indicator like life expectancy, progress internationally appears to be roughly correlated with performance in post-conflict transition, as with under-five mortality. Trends in primary school enrollment do not, however, seem to behave this way, with steep increases in enrollment for countries in all three performance clusters. Life expectancy in Afghanistan has been rising steadily (figure 1.3), at a rate similar to middle performers but slower than strong performers, but starting from a very low base in 2001. At 48 years it remains well below the averages for comparator groups—indeed, below most countries except some in Sub-Saharan Africa affected by high rates of HIV/AIDS.3 Internationally, life expectancy increased significantly on average for strong performers and middle performers, but stagnated or declined for weak performers. 3 The 2010 Afghanistan Mortality Survey shows dramatic improvements in life expectancy and under-five mortality (Afghan Public Health Institute and Central Statistics Office of Afghanistan 2011). These data are not included in the figures because they do not have time-series data for the last 10 years, and for the figure’s international comparative analysis it is essential to use statistics from a single international database (in this case the World Development Indicators). The new 2010 estimates would, if they are confirmed, put Afghanistan in the strongly performing group of countries. 12 Figure 1.3 Changes in life expectancy (top) and under-five mortality (bottom) 75 70 65 60 Life expectancy 55 50 45 40 35 30 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Years after start of transition 240 220 200 180 Under-five mortality per 1,000 live births 160 140 120 100 80 60 40 20 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Years after start of transition Source: Data from World Development Indicators; figures constructed in Byrd (2012). Note: Life expectancy in years at time of birth, under-five mortality per 1,000 live births (on vertical axes); horizontal axes represent years after start of transition (2001 for Afghanistan). Key is from figure 1.1 and is not repeated in subsequent figures. 13 Afghanistan has only slowly reduced the under-five mortality rate over the past decade according to the World Development Indicators (see figure 1.3, and also the discussion in footnote 6). This is in contrast to the strongly performing comparator countries, which generally had sharp declines, whereas middle performers had a slowly declining trend (but better than Afghanistan’s). Weak performers had a rising trend. Thus Afghanistan has been falling further behind the better-performing comparator countries for under-five mortality. Governance and rule of law Restoration of security and rule of law are important elements of transition, and strongly performing comparator countries show a clear upward trend in the rule of law indicator of the Worldwide Governance Indicators (WGI; figure 1.4). Rates of improvement are on average slower for the other two clusters, and among weak performers most countries saw declines or at best stagnation in this indicator. For Afghanistan, levels and trends for the rule of law indicator are poor, with the country starting at one of the lowest levels of any comparator and, after an initial rise, declining from 2005. The World Bank’s World Development Report 2011 underlines the importance of rule of law. The report’s central message is that strengthening legitimate institutions and governance to provide security, justice, and jobs is crucial to break out of cycles of violence in countries suffering from state fragility. The rule of law relates to two of these three priorities. The WGI government effectiveness indicator shows a clear upward trend with some fluctuations for the strong performers, no clear trend for middle performers, and a slight downward trend for weak performers, as well as significant declines for some countries. For Afghanistan this indicator first rose and then declined starting in 2005 (see figure 1.4). Both the starting point and latest observations for Afghanistan are well below the averages for all comparator groups (and below all individual country observations beyond a year or two after the start of transition, except for the Democratic Republic of Congo). Afghanistan’s recent performance puts it near the bottom of the poor performers, but it seems unlikely that overall government effectiveness was much worse in 2009 than in 2003 or 2004, when government functions were just restarting. Moreover, the negative trend in this indicator contrasts with evidence of steady progress in improving concrete elements of government effectiveness in Afghanistan—such as public financial management and effective service delivery under national programs, most notably for health and community development (see chapter 5). This all suggests that trends over time (particularly the reversal starting in 2005) tend to be more reliable for these kinds of perceptions-based indicators than their absolute levels. 14 Figure 1.4 Deteriorating rule of law (top) and government effectiveness (bottom) indicators 1.5 1.0 0.5 0.0 WGI rule of law indicator -0.5 -1.0 -1.5 -2.0 -2.5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Years after start of transition 0.5 0.0 WGI government effectiveness indicator -0.5 -1.0 -1.5 -2.0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Years after start of transition Source: Worldwide Governance Indicators; figures compiled in Byrd (2012). 15 Political stability Afghanistan’s deteriorating political stability in recent years (as measured by the WGI political stability indicator) may be one of the central threats to a successful transition. International experience summarized in World Bank (2011) underlines the importance of political stability (based on “inclusive- enough” coalitions) as the basis for successful transitions from protracted cycles of violence. (Wider political economy considerations are reviewed below.) But Afghanistan’s political stability rating has fallen since 2005, and by 2008–10 it fell lower than that of virtually any comparator country in any year—and considerably lower than any comparator at the corresponding stage in transition (figure 1.5). This contrasts sharply with strongly performing comparator countries, which showed a strong positive trend amid fluctuations and variations. Middle and weak performers had more mixed trends, showing average improving political stability, albeit at lower levels than for strong performers. Figure 1.5 Deteriorating political stability indicator 1.5 1.0 0.5 0.0 WGI political stability indicator -0.5 -1.0 -1.5 -2.0 -2.5 -3.0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Years after start of transition Source: Worldwide Governance Indicators; figures compiled in Byrd (2012). Perceptions-based indicators like the WGI political stability indicator are imprecise and need to be interpreted with caution. For example the expanding Taliban insurgency, not just internal political issues, may have been a major factor contributing to deterioration in this indicator for Afghanistan. But trends in comparator countries and Afghanistan’s relative decline are consistent with evidence from country case studies suggesting that political stability is crucial for successful transition. 16 Unpublished analysis for the World Development Report 2011 of Cambodia, Mozambique, Rwanda, South Africa, and Vietnam—five countries with different political systems that made the transition from conflict to stability and development—suggests four important factors in their success (Byrd 2012). First, the political dimension’s centrality was clear-cut. Many successful comparator countries began their transitions facing political instability, but a dominant party emerged that was able to reinvent itself as inclusive and resilient by consolidating power through deepened authority, managerial executive capability, privileged patron-client networks, and effective use of international aid (despite aid effectiveness and dependency issues). In the case of Afghanistan, the Bonn Agreement of December 2001 outlined a political process with steps leading to a new constitution and elections—and with minor exceptions those steps were taken—but the Taliban were not part of this process. After the completion of the Bonn framework with national elections in 2004 (Presidential) and 2005 (Parliamentary and Provincial Councils), the expanding insurgency as well as other problems hindered further progress in political consolidation. Second, local actors dominated, and were able to take control of key arenas (such as the security sector, revenues and aid) but exploit them for national political consolidation and development rather than narrow factional purposes. They elaborated a positive narrative and gained a dominant negotiating position internally and externally in terms of security, stability and development, inclusivity, nationalism and international respect. These aspects are likely to be very important in Afghanistan as well, highlighting the need for strong Afghan leadership to pursue and protect a national agenda. Third, civilian control of national security forces was integral to a successful transition. Although Afghanistan has a formal system that clearly grants civil authorities oversight over the ANSF, vigilance will be needed during transition to ensure that the ANSF remains firmly under civilian control and that local militias’ primary loyalties are to central authority rather than regional strongmen. Fourth, these countries’ experiences demonstrate that ethnic and other divisions, even if exacerbated during conflict, need not be an insurmountable obstacle to transitioning toward political stability; on the contrary, progress in addressing such issues can become part of the overall process. Moreover, recent comparative research suggests that political institutions—formalized political parties that have an existence and agenda beyond individual personalities—is a significant positive factor contributing to beneficial development outcomes as well as political stability and avoidance of conflict (Keefer 2008 and 2010). In Afghanistan incentives to form and perpetuate institutionalized political parties appear, however, to be weak. This reflects its difficult history during the Communist and mujahideen eras and the adoption of the single nontransferable vote system for parliamentary elections, which strongly favors individual local power-holders rather than consolidation of political parties.4 Transitions also fail, the World Development Report (WorldBank, 2011) discusses the many countries that go through cycles of violence and civil war. Slipping backwards is easy, particularly where violence is a normal part of political negotiations and factions make short-term political deals—and then unscramble them, leading to endemic violence and a repeat of the cycle. Sudan provides a clear example (see Byrd 2012 and De Waal 2009) and offers a lens to view Afghanistan’s experience, including how temporary and shifting deals among different factions and actors, supported by various forms of financial patronage (including aid), have been the norm. 4 See Keefer (2011), and numerous sources cited therein as well as other sources. 17 Political economy The political economy impact of transition for Afghanistan is likely to be at least as important as the direct economic effects. The legacy of 1978–2001 includes chronic violence; ethnic fragmentation; uncertainty and insecurity; the rise of regional and local strongmen; heavy reliance on ethnic, community, and kinship groups for survival and support; and, in a weak economy with few other significant income streams, a substantial opium trade and other illicit activities. The post-2001 period ushered in major changes but also has strong elements of continuity with the past. Critical aspects of Afghanistan’s political economy are:  Uncertainty over longer term international engagement with Afghanistan and future political developments, including the future course of the conflict and the chances of a peace settlement.  Considerable insecurity in many parts of the country, including an active insurgency but also local conflicts, criminality, and predation by criminal groups.  Ethnic and factional fragmentation and associated tensions, and a lack of solid and durable coalitions of common interest.  Large financial inflows (military and civilian) and opportunities for various Afghan actors to benefit from them. These inflows have become a major source of rents, patronage, and political power, and sometimes inadvertently worsened conflicts by affecting the strengths and interests of various elite groups (such as military and political leaders from different factions) and technocrats (Suhrke and Hakimi 2011). These factors can encourage short-term-oriented economic and political behavior; widespread corruption and predatory activities; factionalism (sometimes manifested in armed violence); continuing reliance on armed groups and local or regional powerbrokers; and a focus on (and conflicts over) capturing benefits from financial inflows, particularly through connections involving formal government positions and political office. The political economy ramifications of declining aid and international military spending could be large in Afghanistan, and the transition may exacerbate some of the above features, especially if poorly managed. Elite groups’ adaptations to declining aid and associated realignments and dislocations could be violent and even destabilizing. As international resources fall, reliance on the opium economy and other illicit activities may rise, at least in relative terms, putting greater pressure on central state authority. And political economy pressures are likely to grow on other, less vulnerable funds, such as domestic revenue and budget spending. Mining revenues, if they materialize as hoped (see boxes 2.2 and 2.3) will also carry political economy risks, particularly if they are badly managed or not subject to robust governance and transparency arrangements. Over the longer term, however, less aid could have positive political economy implications because elite groups will have fewer resources to contest and fewer opportunities to egregiously accrue financial benefits. So the challenge will be to manage the short-term risks while striving to exploit the opportunity to promote more positive political economy outcomes over the medium term. Possible actions include the following:  Reducing uncertainty about future international engagement: credible commitments on the size and shape of the international support to Afghanistan beyond 2014 would help lengthen time horizons and mitigate short term behavior.  Improving aid effectiveness by channeling more aid through the budget: this may help reduce the political economy distortions that fragmented and off-budget aid inadvertently encouraged. 18  Maintaining and further strengthening public financial management systems and protections, and generally supporting a robust Afghan budget process: this will strengthen the country’s ability to resist political economy pressures and to support national political agendas and priorities.  Improving the process for government appointments: even if a fully merit-based system will take time to take hold, the appointment process should be more inclusive and evenly spread across different groups and factions.  Ensuring that remaining international contracts (military and other) during transition are spread more broadly: efforts to do so are already under way by NATO and the United States.  Continuing the international support for and monitoring of electoral processes in Afghanistan: though the objectives must be modest, this can help limit violence and other irregularities at the polls. Conclusions and recommendations Afghanistan is moving into the transition under a unique set of circumstances. Not all the lessons from other countries or even Afghanistan’s own history are directly applicable, but they can provide some idea of the processes and factors that the country will face in coming years as foreign troops withdraw and international attention wanes. One of the main lessons from other countries is that successful transitions require political consolidation and strong national leadership. Political uncertainty and insecurity could undermine Afghanistan’s transition and development. International experience demonstrates that violence and especially protracted internal insurgency are extremely damaging to development, and that political stability and consolidation are key ingredients of transitions to peaceful development. This underlines the importance of reaching a peaceful solution to the Taliban insurgency, and the need for political consolidation particularly in the run-up to the next election cycle. (The presidential election is to be held in 2014, and parliamentary elections in 2015.) But if there is worsening insecurity and increasing uncertainty about longer term stability, Afghanistan’s development prospects will be harmed. Political consolidation is not the only pressing issue that needs to be tackled. Afghanistan’s political economy over the past 10 years has been shaped by large inflows of aid, which provided benefits to various groups from the associated rents and contracts. Given these distortions, returning to a more “normal” economy with much lower aid inflows will not be easy and will require continued and concerted international support. Afghanistan’s political elite will need to rise above short-termism, factionalized politics, widespread patronage, and corruption and more consistently pursue a medium- to longer term national agenda. This will require political consolidation and a broad-based coalition for peace and reconciliation. And the international community will need to move beyond viewing transition just in security terms, as a way to exit, and start seeing it as an opportunity to enhance the coherence and effectiveness of its assistance and pursue realistic developmental objectives over the longer term. There is much that both partners can do to make transition more likely to succeed from the broader and longer term perspective taken here. While challenging, the fiscal, government capacity, service delivery, economic, and poverty dimensions of transition can be managed, provided that the overall security and political context of transition (including the regional environment faced by Afghanistan) is conducive. The following chapters examine the impacts of these various key economic, financial, and public sector dimensions of transition as Afghanistan moves through and beyond transition. 19 References Afghan Public Health Institute and Central Statistics Office of Afghanistan. 2011. Afghanistan Mortality Survey 2010. Kabul. Byrd, William. 2012. “Afghanistan’s Transition in International and Historical Perspective.” Background Paper, World Bank, Washington DC Byrd, William, Gary Milante, and Kenneth Anye. 2012. “Afghanistan Transition Study: Quantitative and Indicator-Wise International Comparisons.” Background Paper, World Bank, Washington DC. De Waal, Alex. 2009. “Mission Without End? Peacekeeping in the African Political Marketplace.” International Affairs 85 (1): 99–113. Fotini, Christi. 2011. “Lessons from the Soviet Withdrawal for the Present-day Afghan Transition.” Background paper, World Bank, Washington DC. Keefer, Philip. 2008. “Insurgency and Credible Commitment in Autocracies and Democracies.” World Bank Economic Review 22 (1): 33–61. ———. 2010. “Why Follow the Leader? Collective Action, Credible Commitment and Conflict.” Prepared for the Oxford Handbook of the Economics of Peace and Conflict, eds. M. Garfinkel and S. Skaperdas. ———. 2011. “Political Obstacles to Growth and Public Service Delivery in Afghanistan.” Background paper for Afghanistan growth study, World Bank, Washington DC. Menkhaus, Ken. “Somalia and the Horn of Africa.” WDR 2011, Background Case Study. Suhrke, Astri. 2006. “When More Is Less: Aiding Statebuilding in Afghanistan.” Working Paper 26, Fundación para las Relaciones Internacionales y el Diálogo Exterior, Madrid. Suhrke, Astri, and Hakimi, Aziz. 2011. “Sustainability and Transition in Afghanistan. A Political Economy Analysis.” Background paper, World Bank, Washington DC. World Bank. 2011. World Development Report 2011: Conflict, Security, and Development. Washington, DC. 20 Chapter 2 The Economic Impacts of Transition Summary Aid has had a large impact on Afghanistan’s economic growth. Depending on future aid flows, GDP growth will drop from 9 percent in 2010/11 to closer to 5 percent on average until 2018, and will likely be even lower in the long term. Given current high rates of population growth (2.8 percent a year), economic growth at these levels will be too slow to reduce the numbers of people living in poverty over at least the next 15 years. Employment and poverty are unevenly distributed across the country, and both are somewhat sensitive to inflows of aid, as beneficiaries tend to live in conflict-affected provinces and better-off households. Thus the largest impact of declining aid on employment and (to a lesser extent) poverty is likely to be felt in conflict-affected provinces, leading to a loss in household income and fewer opportunities for casual labor and generally lower wages. But the overall impact of declining donor assistance on household welfare is limited by the current focus on security in public spending. A gradual fall in aid might be beneficial in the long term as it would reduce distortions in the economy caused by the extraordinarily high levels of aid in the past. A rapid decline could, however, lead to major macroeconomic instability and serious socioeconomic consequences. Future aid flows need to be carefully programmed to allow the economy to adjust to new opportunities for growth from mining (especially), agriculture, and services. Making aid more effective can mitigate some of the negative impact of the fall in aid. Raising the local content of aid could do this, through shifting more aid on budget and increasing opportunities for local vendors to participate in aid contracts. Rather than declining aid, however, deteriorating security and governance are likely the largest risks to the economic outlook. Efforts to sustain or improve security—as well as to bolster the legitimacy of the state—will remain of utmost importance. Introduction Afghanistan is a highly aid-dependent country. Such aid dependency is almost uniquely high internationally, and has important implications for the macroeconomic environment and socioeconomic indicators. After a decade of internationally supported reconstruction efforts and security assistance, the country now faces a transition involving withdrawal of most international military forces by end- 2014. International experience shows that such withdrawals often prompt declines in aid. Given Afghanistan’s high aid dependency and relatively narrow economic base, this raises important questions about the transition’s economic effects. 21 The economy Afghanistan is one of the poorest countries in the world. In 2010 GDP was US$ 15.9 billion (excluding opium production), after nearly a decade of strong growth. With an estimated population of 30.6 million, the country has a per capita GDP of US$ 528. In spite of encouraging progress, social indicators are still dismal (and suggest wide gender gaps): 36 percent of people are poor and about three-quarters of the population are illiterate. Since the fall of the Taliban in 2001, the country has seen exceptionally high economic growth. Real GDP grew at an average annual 9.1 percent from 2003/04 to 2010/11 (figure 2.1). High growth rates are not unusual for post-conflict countries. Conflicts often result in massive physical and human capital destruction, and during the recovery phase, investments—often financed by foreign aid—generate high returns to capital accumulation. This often leads to higher-than-average growth immediately post- conflict. Although the economic performance of post-conflict countries displays considerable heterogeneity, Afghanistan’s has been much stronger than many (chapter 1). Figure 2.1 Real and agricultural GDP growth since 2003/04 (percent) 50 40 30 21.0% 20 13.7% 11.2% Percent 10 8.4% 8.4% Real GDP growth 5.6% 3.6% 1.1% 0 -10 -20 Agriculture GDP growth -30 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11* Source: World Bank staff calculations. Note: Excluding opium production. Total military and civilian aid in 2010/11 was US$ 15.7 billion—about the same as GDP, reflecting sharp increases since 2005. Foreign aid disbursements (security-related and civilian) were equivalent to nearly 100 percent of GDP in 2010/11.5 Foreign aid has been on the rise since 2002, growing from US$ 404 5 Aid refers here to normal civilian aid as well as security-related aid (the larger component), but excludes the huge amounts spent on international military forces. 22 million in 2002/03 to more than US$ 15.7 billion in 2010/11 (figure 2.2). Aid typically translates into demand for domestic services and construction through local procurement of goods and services by donor and military agencies. Figure 2.2 Aid trends, Afghanistan Source: World Bank staff estimates. It is therefore unsurprising that services have been a strong driver of the economy over the last decade (figure 2.3). In 2010/11 they constituted 51 percent of GDP (table 2.1) and contributed 9 percentage points of the 8.4 percent real GDP growth (figure 2.4). Agriculture contracted in 2010/11 due to bad weather, lowering overall growth by 1.5 percent. Agricultural value added has declined as a share of GDP over the last decade. Industry accounts for 26 percent of GDP: construction growth is still relatively strong, feeding growth in manufacturing through demand for building materials. Contributions from mining are almost negligible, but are expected to surge over the next decade. 23 Figure 2.3 Sector growth rates, 2003/04–2010/11 80 70 60 50 Percentage of GDP 40 30 20 10 0 -10 -20 -30 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 Agriculture Industry Services Source: IMF and CSO, World Bank staff calculations. Note: Excluding opium production. Table 2.1 Sector shares in GDP, percent Sector 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 Agriculture 38.5 36.8 28.5 30.2 27.2 27.3 22.4 27.0 23.3 Mining 0.1 0.1 0.2 0.3 0.3 0.5 0.5 0.4 0.6 Manufacturing 18.7 18.1 18.8 17.4 17.1 15.6 15.4 13.5 13.0 Construction 4.8 6.0 9.4 11.2 13.8 13.2 14.0 12.5 12.4 Other industry 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 Services 37.8 38.8 43.1 40.9 41.6 43.4 47.6 46.5 50.7 Source: IMF and CSO, and World Bank staff calculations. Note: Excluding opium production. 24 Figure 2.4 Sector contribution to GDP growth (percent) 25 20 15 Percentage of GDP 10 5 0 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 -5 Services Other industry Construction Manufacturing Mining and quarrying Agriculture -10 Source: IMF and CSO, World Bank staff calculations. Note: Excluding opium production. The fastest-growing subsectors in services since 2003/04 have been communications (66 percent annual average), financial and business services (26 percent), government services (20 percent), and transport (14 percent). The impact on services growth of financial and business services is still marginal (considering the subsector’s growth and share in total services). The large share and strong growth of communications and transport, followed by government services, have been main drivers of overall economic growth. These subsectors, along with construction, are typically driven by the needs of post- conflict reconstruction and high inflows of aid. Wholesale and retail trades also play an important role, but mainly through their weight in total services rather than strong growth. Growth in Afghanistan is volatile, mainly due to the economy’s heavy reliance on agriculture, which in 2010/11 contributed 23.3 percent of total value added. Agricultural production is particularly susceptible to the weather. During 2003/04–2010/11 annual real agricultural growth ranged from 22 percent contraction to 45 percent expansion (see figure 2.3), and the sector accounted for between a quarter and a third of GDP. Absorbing 60 percent of the working population, agriculture is by far the largest employer, so rural development will remain crucial in sustaining growth. 25 Agriculture will remain a very important source of income and growth for many years. The illicit production of opium still overshadows agriculture, accounting for nearly half of overall agricultural production but on a much smaller portion of arable land (box 2.1). According to 2011 estimates from the United Nations Office on Drugs and Crime, the farm gate value of opium is 8 percent of GDP, and its overall contribution to GDP (including distribution) is believed to be around a third. Box 2.1 The opium economy Opium is Afghanistan’s most important agricultural crop by value and provides much-needed livelihoods for many people in rural areas, but it distorts incentives to develop a sustainable formal agriculture sector in the long run. In addition, the large criminal profits of the drug industry undermine governance, fuel corruption, nurture dysfunctional politics, and ultimately stimulate insecurity and conflict. Opium cannot therefore be regarded as a sustainable source of long-term growth. Moving away from economic reliance on opium is thus a priority development objective. But given opium’s characteristic as both a high-value, storable commodity with a ready market and a secure cash crop for an insecure environment, this will not be easy. Still, progress in eliminating opium cultivation in some areas, which has been sustained in some cases for many years, suggests that phasing out opium production in Afghanistan over the next 10–20 years is possible. Estimates for the area planted to opium poppy have fallen sharply in recent years, from 193,000 ha in 2007 to 131,000 ha in 2011. The decline was largely driven by saturated markets and lower prices, even though trends are rather volatile. To some extent government efforts to restrict opium cultivation and trade played a role as well, resulting in an increase in the number of “poppy-free” provinces and leaving the opium economy highly concentrated in the insecure southern part of the country. Opium production grew in 2011 after a plant disease wiped out nearly half of the harvest in 2010. This increase was compounded by a drop in prices for wheat (income from opium was 11 times higher than that from wheat in 2011). It is conceivable that during the transition opium production will grow temporarily, because some households may increase opium production as a coping mechanism in the transition to make up for the loss of income from other economic opportunities during the transition process. Mining makes a marginal contribution to GDP—less than 0.5 percent during the 2000s—but could potentially take off in the future. Afghanistan has substantial untapped mineral deposits (box 2.2). Two large investments in copper deposits in Aynak (near Kabul) and iron deposits in Hajigak (Bamyan province), in 2010 and 2011, could be the first step toward realizing this potential. Box 2.2 Mining as a potential source of growth Afghanistan has copper, gold, iron ore, and other minerals, as well as construction materials, dimensional stone and gemstones, coal, and hydrocarbons (mostly natural gas). Some of these resources were exploited in the 1970s and 1980s, but war, neglect, and paltry funding in the last two decades caused output to plummet far below potential. Mineral production is now limited to small coal operations, limestone, construction materials, and gemstones and dimensional stone. This all could change dramatically over the next few years. Mining is seen by many observers to have high potential, and some state that it will become important for growth in the near future and will lead economic growth over the longer term. Similar sentiments place very large numbers on the value of Afghanistan’s mineral resources—often well over a trillion dollars—for a country with a GDP in 2010 of only US$ 15.9 billion. These “projections” have some element of truth, but refer mainly to the value of minerals in the ground: they assume that the minerals are in large enough concentrations to be mined profitably and that the necessary infrastructure will either be available or can be built profitably to extract and sell them. Such prognoses may hold in the long run, as mining techniques improve, resources get scarcer, and the economy develops. But in the very short run only two deposits are reasonably sure of being mined profitably: the copper deposit at Aynak, just 35 kilometers south of Kabul, and the ore deposit at Hajigak, in Bamyan province. 26 Partial equilibrium analysis shows that these two mines have the potential to generate substantial economic returns. The revenue from these two mines is estimated at US$ 322 million a year in 2011–15, primarily in signing bonuses, and over US$ 900 million per year on average until 2031. Opening 11 other mines (box 2.3) has the potential to increase revenue to around US$ 1 billion a year on average from now through 2031. Mining could eventually contribute 2–2.5 percent of GDP to domestic revenue (which stood at 11 percent of GDP in 2010/11) and add 2–3 percentage points to annual GDP growth. But the direct employment effect is expected to be relatively modest. Aid dependency Afghanistan is an extreme outlier in terms of dependence on aid (figure 2.5). While the bulk of aid is security related, civilian aid is estimated at more than US$ 6 billion a year (including on-budget assistance), or nearly 40 percent of GDP. Donor grants accounted for 50 percent of the core budget in 2010. Off-budget contributions are estimated at US$ 9.4 billion for the same year. Such aid dependency is almost uniquely high globally (only a few much smaller entities, such as Liberia and West Bank and Gaza, receive more aid per capita). Understanding how aid affects the economy is critical to appreciating the economy’s dynamics and the likely impact of the transition. Figure 2.5 Aid versus GDP per capita, average of 2006–08 100 90 Net official development assistance received (percent) 80 Afghanistan 70 60 50 40 30 20 10 0 4.5 5.5 6.5 7.5 8.5 9.5 GDP per capita (current $) Source: World Bank staff calculations. Note: Aid includes security-related aid. 27 Aid is delivered through three basic modalities (table 2.2). First, in 2010 the government received around US$ 1.9 billion as on budget support. Some of it (US$ 780 million) finances the salaries of the Afghan National Army and Afghan National Police. The rest finances development projects and some recurrent government expenses. Second, donors finance development projects through external budget contributions (US$ 5.2 billion in 2010), administered by civilian aid agencies such as USAID, the Department for International Development of the United Kingdom, and the European Commission. Third, donors provide security-related aid, also outside of the budget, to finance operations and investments of the International Security Assistance Force through military bodies and contracting agencies, such as the U.S. Corps of Engineers and provincial reconstruction teams (PRTs).6 This is the largest component, coming to US$ 8.6 billion in 2010. Table 2.2 Donor assistance (US$ million, unless otherwise indicated) Donor assistance 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 Civilian aid 2,416 1,350 2,188 2,675 3,942 5,262 Security-related aid 988 1,905 7,028 2,750 5,470 8,594 On-budget support 720 717 1,069 1,024 1,275 1,886 Total 4,124 3,972 10,284 6,449 10,686 15,742 Percentage of GDP 66 56 118 61 86 98 Source: World Bank staff calculations. Note: The figures reflect actual disbursements for the respective year and are based on self-reporting by the largest donors. Civilian aid includes expenditure of the U.S.-financed Commanders Emergency Response Program. Data is cross-tabulated with various sources. Aid inflates the government’s role in the economy. In 2006/07–2008/09 government consumption as a share of GDP was 10.8 percent, which is slightly lower than expected for the country’s income. But those data do not include public aid projects outside the core budget—that is, externally budgeted—which in Afghanistan are exceptionally high. If, therefore, we define public consumption as the sum of government consumption through the core budget and through the external budget—usually considered part of private consumption—the proportion of public consumption in GDP climbs to 47.5 percent, making Afghanistan an extreme outlier in that ratio too (figure 2.6).7 6 PRTs are civil–military teams that operate in most provinces under the command of a lead nation. They are a wide channel for the off-budget Commanders Emergency Response Program and bilateral donor funds to the provinces and generally execute their development activities outside of and parallel to those of the government. 7 Consumption represents 47.7 percent of the core budget and 56.5 percent of the external budget. 28 Figure 2.6 Government consumption versus GDP per capita, average of 2006–08 60 General government final consumption expenditure (percentage of GDP) 50 Afghanistan, including off-budget public consumption 40 30 20 10 Afghanistan 0 4 5 6 7 8 9 10 11 12 GDP per capita (current $) Source: WDI, World Bank. Public investment (17.8 percent of GDP in 2010/11) is much higher than private investment (10.6 percent, including foreign direct investment) and both have declined since the mid-2000s. Public investment as a share of GDP is similar to private investment levels in other countries, again mirroring the country’s aid-dependent economic structure and highlighting the need to reverse these shares. Spending “on” Afghanistan does not equal spending “in” Afghanistan. Cumulative U.S. spending for the Afghanistan mission is estimated to be as high as US$ 444 billion. The U.S. Department of Defense appropriated US$ 118.6 billion alone in 2011 (Belasco 2011). The sheer size of U.S. spending on Afghanistan has led many to assume that military withdrawal will have a very negative impact on the economy. However, military spending by the United States (and other countries) finances the salaries of military personnel, investments in weapons equipment and systems, sustainment, logistics and research of international forces, and operations contracted and paid for outside the country. Although it indirectly benefits Afghanistan’s economy by supporting security, the direct positive impact on poorer households appears to be limited. The impact of its withdrawal is therefore likely to be muted. Most aid (both civilian aid and security assistance), including the amount contracted in-country, has a low domestic economic content, limiting its impact on the economy. Much either never comes in or flows directly out through contracting international providers of goods and services, imports, and the expatriation of profits. A study by the Peace Dividend Trust (2008) suggests that the domestic content from trust funds and budget support is around 70–80 percent, and only 35–50 percent and 10–20 percent respectively for local and international contracts, resulting in an overall local impact from donor aid of 38 percent. In other words, only 38 cents of every aid dollar spent in Afghanistan actually reaches 29 the economy through direct salary payments, household transfers, or purchase of local goods and services. This study, however, did not consider security spending. It is estimated that the domestic content of aggregated aid flows is only about 14–25 percent (table 2.3): US$ 2.2–US$ 4 billion of aid may have stayed in the country in 2010/11. To calculate this figure, the methodology applied by the Peace Dividend Trust (2008) was modified and used a combination of data collection, focus group interviews, and assumptions.8 Table 2.3 Domestic economic content of aid, 2010/11 (percent) Aid On budget External budget Security-related 90–95 10–15 Civilian 55–60 17–25 Source: World Bank staff calculations. Impact of transition on economic growth Given Afghanistan’s high aid dependency, how will the decline of aid affect economic growth? To tackle this question, we ran a set of simulations in a dynamic computable general equilibrium model that mimics the characteristics of Afghanistan’s economy. The simulations allow us to analyze alternative aid and development scenarios and their impact on economic growth and economic variables (box 2.3). This section reports only selected results. Appendix 1 provides a detailed description of the model, its parameters, and the assumptions for the scenarios discussed here. Box 2.3 Overview of scenarios and assumptions For this report, we analyzed six different scenarios, summarized as follows:  Gradual decline in aid (BASE). This scenario represents a reasonable development outcome for Afghanistan and is the benchmark for the other scenarios. Its basic assumption is that the main variables of the economy develop in a manner that is similar to recent trends with little structural change, except for the development of mining. It assumes, in particular, a gradual decline of external aid (security and nonsecurity) and on-budget aid resulting in total aid of about 23 percent of GDP in 2018 and 10 percent in 2025. Private and public demand (consumption and investment) shares are around 49 percent and 51 percent; this scenario assumes that they gradually improve to 75 percent and 25 percent. Other government revenue effects include an assumed increase in the tax take. Mining development reflects expected developments for Aynak and Hajigak only. Opium production, as part of the agriculture sector, is assumed to stay constant in the model.  Limited agricultural growth (AGR-). This illustrates the sensitivity of the BASE scenario. Unlike BASE, which assumes average agricultural growth of 6.3 percent annually, this scenario assumes agricultural growth closer to the historical average of 3.9 percent.  Improved security situation (MIN+). Departing from the BASE case, this scenario assumes exploitation of 11 other mines (in addition to Aynak and Hajigak) due to higher private investment encouraged by improved security.  Deteriorating security situation (GOV-). This scenario represents the consequences of a worsening security and governance situation with substantially lower levels of aid, lower investment, faster depreciation of capital investment, and nearly no growth in productivity. 8 Due to lack of data, we were unable to update the local content ratio for off-budget civilian aid. However, the partial data (collected from a smaller sample of aid agencies) suggest that local content has stayed much the same as in 2006. 30  Rapid decline in aid (AID-). Under this scenario, donor-managed (off-budget) grant aid starts to decline faster from 2012, reaching 14 percent of GDP by 2018 and then gradually decreasing to 10 percent of GDP in 2025. Government on-budget aid is unchanged from the BASE scenario.  Shifting more aid on budget (AIDALLOC). This scenario makes the same assumptions as BASE except that 50 percent of all aid is channeled through the budget starting in 2014/15. The model indicates that in a scenario with a gradual decline in aid (BASE), GDP growth (at factor cost) may drop from 9.1 percent a year in 2003–10 (8.4 percent in 2010/11 itself) to 5.9 percent until 2018/19 (table 2.4). Donor consumption and investment decrease, and government expenditure stays more or less the same with a slight shift from investment to consumption. Imports decline as expected while aid is phased out; exports increase, not least because of the start of mining output but also due to the depreciated value of the Afghani. After 2018 until 2025, GDP growth may drop further to around 3.5– 4 percent. Table 2.4 Real macro indicators by simulation, 2010/11–2018/19 (percentage annual growth) BASE AGR– AIDALLOC AID– GOV– MIN+ GDP at factor cost 5.9 4.9 6.1 5.2 0.7 7.0 Consumption Private 5.7 4.6 6.0 5.0 1.8 5.9 Donors –14.1 –13.9 –21.3 –23.5 –56.7 –14.1 Government 1.6 2.0 7.4 1.6 –4.9 2.6 Fixed investment Private 3.5 2.4 4.3 2.8 –14.7 10.3 Donors –10.2 –10.1 –17.7 –16.0 –63.3 –10.2 Government 1.3 1.2 9.8 1.0 –7.9 2.6 Exports 10.1 8.9 9.2 9.6 4.3 12.5 Imports –2.2 –2.9 –2.8 –4.3 –10.7 –1.2 Absorption 0.1 –0.7 0.3 –1.4 –6.4 0.7 Real exchange rate (index) 2.8 3.4 2.3 2.9 1.8 2.8 Source: World Bank staff calculations. Mining is expected to contribute significantly to growth, accounting for 2.8 percentage points of the 5.9 percent total average growth in the BASE scenario (figure 2.7). Assuming additional investment in mining (MIN+), due, for example, to an improved security situation, mining could contribute even more (4.8 percentage points if a further 11 mines are opened) and increase medium-term growth to 7 percent on average. Agriculture remains an important source of growth, not least due to its high share of GDP. Depending on the weather and investment, average GDP growth could produce volatility of 5–6 percent annually on average. But the contribution of services, which accounted for 9.2 percentage points of growth in 2010/11, is set to fall heavily and account for an average of only 0.3 percent of growth under the BASE scenario, mainly reflecting weaker demand for transport and retail services. 31 Figure 2.7 Decomposition of GDP at factor cost by activity and simulation, 2010/11–2018/19 (percent) 8.0 6.0 4.0 Percentage of GDP 2.0 0.0 BASE AGR– AIDALLOC AID– GOV– MIN+ -2.0 -4.0 -6.0 Agriculture Opium Mining Industry (nonmining) Services Source: World Bank staff calculations. Shifting more aid on budget (AIDALLOC) could increase average annual growth from 5.9 percent (in BASE) to 6.1 percent. The difference is explained by a higher contribution from services. Shifting aid on budget has therefore a slight mitigating effect on growth as the local impact content is larger for on- budget than off-budget spending. While the difference appears marginal, other economy-wide effects— for example, more local employment—could be important. A more rapid decline in aid (AID-) than in the base scenario could lead to a decline of GDP growth of nearly 1 percentage point to an average of 5.2 percent (and to 4.7 percent by 2018/19). Transmission would mainly be through a larger negative effect on services growth. All these results need to be handled with considerable care. An economic model is only as good as its underlying data, which in Afghanistan are still very weak (box 2.4). 32 Box 2.4 Limitations of the model The model exploits all available macro and fiscal data, sector studies, enterprise surveys, and household survey sources from the government, as well as International Monetary Fund (IMF) and World Bank publications. Whenever necessary, data for comparable countries were used to fill in for missing information. Due to the simplifications that are inherent in any model, combined with the particularly weak datasets for Afghanistan, readers should view the results as approximate indicators of the effects of alternative policy options, to be considered with other inputs to economic policymaking in Afghanistan. Readers should also remember that the model has been developed to assess the integrated effects on an economy of certain policy or external variables. It is a model intended to help the policy discussion in considering the overall general equilibrium effects, not just the direct partial equilibrium effects. Thus the quantitative results and the analysis of changes in parameters, size, and direction should only be taken as an indication of trends. For sector- specific, year-to-year developments, more detailed sector studies and fiscal sustainability analyses are required. Finally, the model hinges on a set of underlying assumptions about developments in growth in Afghanistan that depend on a stable security environment —or at least on a general improvement in the investment climate (due to a stable security environment, institutional reforms, and infrastructure improvements)—that leads to higher levels of private investment. The results are therefore particularly sensitive to changes in the security environment and exogenous shocks, such as higher fuel prices. Risks to macroeconomic stability Compared with many other post-conflict countries, Afghanistan has kept its macroeconomic environment relatively stable. In 2001 it faced weak or nonexistent institutions, low domestic revenue, volatile aid flows, and high import needs (due to lack of domestic output). In this light, Afghanistan has shown a remarkably solid macroeconomic performance over the last decade. Avoiding very high inflation is a prerequisite for a sound economy that offers good growth and employment prospects, and inflation, though volatile, has been relatively moderate. Soon after 2001 Afghanistan ended the hyperinflation of the 1990s, pushing through a highly successful currency reform, and then made good progress in taking inflation lower despite global food and energy price shocks and supply blockages in neighboring countries, which inevitably exacerbated its own price volatility. Still, macroeconomic stability relied heavily on abundant aid flows. And commonly for a country undergoing reconstruction and recovery, its large material needs skewed the trade balance heavily toward imports. After an initial sharp increase in imports, the trade gap has stayed more or less constant with imports about four times exports, giving a huge trade deficit of 64 percent of GDP in 2005/06– 2010/11. In 2010/11 the current account deficit (excluding grants) was estimated at 40 percent of GDP, and was financed—as in previous years—by grants (figures 2.8 and 2.9), confirming Afghanistan’s aid dependency and raising concerns about the impact of aid falls on the current account balance. 33 Figure 2.8 Balance of payments, 2002/03–2010/11 (percentage of GDP) 80 60 40 Percentage of GDP 20 0 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 -20 -40 -60 -80 Trade balance (goods and services) Capital and financial account Current ccount balance (excluding grants) Grants Overall balance of payments Source: IMF data. 34 Figure 2.9 Money reserves, 2002/03–2010/11 16 14.7 14 13.0 13.1 12.6 12 10 9.7 7.8 8 6.8 6 4.9 4 2.8 2 0 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 Source: IMF data. Note: Figure unit is gross reserves in months of next year imports of goods and services, excluding duty free imports and unrecorded reexports. The transition may therefore have a pronounced effect on the balance of payments. Projections of exchange rate behavior in the different scenarios indicate a substantial depreciation of the local currency. Although capital inflows linked to investments in mining are expected, the decline in aid outweighs those inflows; particularly in the first years after 2014 (mining activities are expected to start only in 2016/17). Afghanistan’s fiscal performance in the past decade has been encouraging. The public financial management system underwent a structural transformation that allowed the budget to grow from US$ 346 million in 2002/03 to US$ 3.3 billion in 2010/11, demonstrating rising absorption capacity for funding channeled through the budget. Domestic revenue grew by 40 percent nominally a year over the same period, from a low base of 3 percent of GDP to 11 percent, thanks to reforms in tax and customs administration. Even though domestic revenue repeatedly exceeded fiscal targets, fiscal consolidation relied heavily on donor financing. The risks to fiscal sustainability will remain high and potentially grow. (As the next chapter argues in detail, expenditure is expected to grow much faster than in the past and much faster than any growth in domestic revenue, which will increase the risks to the stability of the fiscal and macroeconomic framework.) Given the absence of alternative financing mechanisms, donor financing will thus remain crucial to sustaining fiscal and macroeconomic stability. In January 2010 Afghanistan reached the completion point for the Heavily Indebted Poor Country Initiative and thus received debt relief under that and the Multilateral Debt Relief Initiative. The debt relief brought down public external debt from US$ 11.9 billion in 2006 to US$ 1.2 billion by end-March 35 2011. Correspondingly, the debt-to-GDP ratio declined from 19.8 percent in 2008/09 to 8.1 percent in 2010/11. Debt service was equivalent to 1.8 percent of exports of goods and services. Nevertheless, as reflected in the Debt Sustainability Analysis (IMF and IDA 2011), Afghanistan will continue to be unable to incur significant debt, which would put the country at high risk of debt distress, due to both insufficient fiscal space to absorb higher debt service and the general vulnerability of the macroeconomic framework to changes in the security situation and external price shocks. The government has indeed incurred very little debt since 2008, underscoring the importance of substantial and long-term grant financing. Lower aid inflows are expected to have little direct impact on real estate prices, mainly due to the sector’s diversified demand, limited exposure to lending, lack of a mortgage market, and little speculation. But given that values have jumped in Kabul and other provincial cities since 2001, reaching rates that are, in the high-end segment, unaffordable even to wealthy Afghans, they are expected to soften. Weaker demand for high-end property should lead to lower average prices once the effect trickles down to other segments. However, systematic data on the real estate market do not exist-these forecasts are based on information from focus group discussions with property experts and analysis of secondary data sources. Impact of transition on employment The impact of transition on employment is very difficult to assess as labor statistics are scarce. Most data come from only three sources: the National Risk and Vulnerability Assessment (NRVA 2007–08), a national labor survey from 2005, and a survey on urban labor markets conducted in 2007 by the Ministry of Labor and Social Affairs, Martyrs and Disabled. All have major limitations and are inconclusive on the relationship between aid and employment. Poor data are worsened by the general lack of systemic employment information among donors. According to the NRVA, 66.5 percent of those in the working-age population (age 16 and above) participated in the labor market. Only around 6.8 percent of the population could be considered unemployed in 2009 based on the conventional definition of the term. However, the relatively low level of open unemployment is counterbalanced by the severity of underemployment: more than 48 percent of those employed work fewer than 35 hours a week on average. Underemployment is particularly widespread in rural areas and among the self-employed, including day laborers and family workers, who make up about 77 percent of the workforce. Employment patterns vary greatly across provinces, though two main patterns emerge. First, regardless of any apparent relationship with the size of the labor force, unemployment rates in conflict provinces are lower than in non-conflict provinces and, in the majority of them, lower than the national average. Second, only four conflict provinces show both low unemployment and underemployment rates. Conflict provinces seem to exhibit marginally higher underemployment than non-conflict provinces (figure 2.10). A likely explanation for these patterns is that job opportunities created by PRTs, including workfare programs, might have helped change the status of workers previously categorized as unemployed, but many of these jobs are likely to be casual and short-term in nature. A likely explanation for these patterns is that job opportunities created by PRTs, including workfare programs, might have helped change the status of workers previously categorized as unemployed, but many of these jobs are likely to be casual and short-term in nature. 36 Figure 2.10 Unemployment and underemployment by province 20.0 90.0 18.0 16.0 80.0 14.0 70.0 Unemployment Rate 12.0 60.0 10.0 50.0 National Average 8.0 40.0 6.0 30.0 4.0 20.0 2.0 10.0 0.0 0.0 Badakhsh … Herat Khost Kunduz Kabul Kapisa Paktika Logar Wardak Daykundi Farah Kunarha Faryab Ghazni Parwan Nimroz Urozgan Helmand Baghlan Sar-I-Pul Takhar Laghman Panjsher Ghor Zabul Nooristan Paktya Nangarhar Bamyan Balkh Jawzjan Badghis Kandahar Samangan Badakhshan Herat Khost Kunduz Kabul Kapisa Wardak Farah Faryab Paktika Logar Daykundi Nimroz Kunarha Parwan Helmand Ghazni Nooristan Laghman Urozgan Paktya Zabul Sar-I-Pul Balkh Baghlan Badghis Takhar Ghor Bamyan Panjsher Jawzjan Nangarhar Samangan Kandahar Source: NRVA 2007/08. Note: Yellow bars = non-conflict provinces, blue bars = provinces with PRTs, blue-black checkered bars = high conflict provinces, black line = national average. It is unclear how many jobs aid has created, partly because the data are widely scattered and difficult to interpret. Because most aid-financed activities are tendered out, information on recruitment and salaries is held by the winning contractor, who has no obligation to give it to the contracting donor. On the donors’ part, efforts to survey employment across contracts face multiple classification problems as many of the jobs are short term, lasting no longer than the contract. A study by the Peace Dividend Trust (2011), which surveyed 146 local businesses who won aid-financed contracts in 2006–2011, found that the average length of a contract was six months, and that more than half the jobs lasted less. But it also found that efforts to improve local sourcing led businesses to retain some workers they employed: 58 percent of businesses who won an aid-financed contract ended the assignment with more employees than they started with. The study’s authors conclude that the contracts, which had a total value of US$ 1.1 billion, created roughly the equivalent of 118,000 jobs for six months. The Ministry of Finance reported that, in 2010, some 6,647 people working outside the civil service in non-security ministries received salaries or salary top-ups financed by donors. The International Security Assistance Force registers 60,000–80,000 Afghans directly employed through military-related contracting agencies. However, information on the length of employment is not available, and there may possibly be errors due to double counting. USAID, depending on the source, estimates 31,600– 60,000 jobs created through its own contracts, but it is unclear if these numbers include second-round effects (jobs created indirectly through aid money). US-CENTCOM reports that U.S.-financed contracts employed 34,200–78,500 Afghans, including contracts issued by USAID and military agencies.9 NGOs are believed to employ around 16,600 Afghans. So there is much uncertainty even over direct employment through aid contracts. Taking employment numbers reported by aid agencies, the number of jobs created is around 410,000. Using the methodology applied by the Peace Dividend Trust (2011), estimates for jobs lasting six months and financed through off-budget aid range roughly between 312,000 and 620,000, depending on local content assumptions.10 Including the number of beneficiaries in the civil service, this suggests that 6.5– 9 US-CENTCOM reports (July 2008–January 2011). 10 We assumed local content of 0.15–0.25 for off-budget expenditure in 2010. The cost of one job was assumed to be US$ 1,108, following Peace Dividend Trust (2011). 37 10 percent of the working population benefited, in one form or the other, from employment created or sustained by aid in 2010. Due to the nature and type of donor involvement in Afghanistan, much of the security-related and civilian aid has been allotted to construction and services, in particular transport and logistics, retail, and maintenance and repair. In U.S. contractual spending alone, these two sectors account for over 75 percent.11 We therefore expect the employment impact from a decline in foreign spending to be most pronounced in these sectors. What happens to employment when aid declines? There are several possible effects:  Assuming a gradual decline in aid, the employment impact is expected to be limited. Unless aid is phased out very abruptly, the employment impact will be felt slowly. Applying the same methodology as above, a decline of, for example, US$ 0.5 billion in (off-budget) aid would affect 11,000–18,000 jobs (on a six-month basis). This is less than 0.2 percent of the current working population and a number that could be absorbed by other emerging economic activities. The Aynak and Hajigak mines, for instance, are expected to directly produce more than 20,000 jobs annually in 2016–2022, in the same sectors where aid spending is now concentrated.  The public sector will likely absorb part of the employment impact. Contingent on continued budget support and fiscal space, the government’s reform intentions are geared toward expanding the public sector: the headcounts of the Afghan National Army and Afghan National Police are expected to increase by 21 percent from current levels by end-2012, some 10,000 teachers are expected to be hired annually, and the government wishes to attract up to 2,000 skilled Afghans from the “second civil service” (chapter 4) into the regular civil service over the next five years. This could mitigate the negative impact, which will be mainly in the private sector.  As in the poverty analysis below, the employment effects will be uneven among provinces. Because most jobs were created in provinces that received higher aid flows, the conflict-affected provinces will bear the brunt of the impact as the transition unfolds, as off-budget provincial aid allocations will likely decline faster than those through the core budget.  Urban centers are likely to be harder hit than rural areas, because construction and services (the most vulnerable sectors) are more important there: they account for 64 percent of the working population in urban areas but only 24 percent in rural areas (figure 2.11).  The employment impact will aggravate underemployment rather than show through in unemployment. As many jobs created by aid are casual, these will probably be the first ones to go when the international troops withdraw and PRTs are dissolved. But as many people involved in daily labor or the casual food or cash-for-work programs also work in agriculture, the transition will more likely lead to higher underemployment and lower household income than higher unemployment. A similar effect might materialize for skilled, high-wage earners who hold (high- paying) jobs in donor projects. They are more likely to find other jobs in the expanding regular civil service, though at lower wages. 11 Information derived from US-CENTCOM contracting database. 38 Figure 2.11 Distribution of the employed population by economic sector, 2007/08 (percent) National Rural Urban Trade Education 11% Trade Agriculture 2% 8% 9% Public Transport and Education Construction, administration communication 2% Transport and mining, and 2% 2% Trade communication quarrying Public 29% 3% 10% administration Other services 4% 7% Manufacturing Manufacturing Other services 8% 4% 9% Agriculture Construction, 59% mining, and Agriculture quarrying Transport and Manufacturing 67% 8% communication 5% Other services 4% 18% Construction, Education mining, and Public 7% quarrying administration 8% 14% Source: ICON Institute 2009. Impact of transition on poverty Declining donor assistance is likely to lower household welfare, but only a little, partly because current spending focuses on security. Recent poverty data are lacking, but two NRVA surveys in 2005 and 2008 provide some evidence on the relationship between spending in the provinces and poverty. 2005 is chosen as the baseline year, coinciding with the steep climb in aid flows to Afghanistan (particularly from the United States). It is assumed that before 2005 (when aid to Afghanistan was generally low), all provinces had roughly similar levels of per capita spending. It is further assumed that any variation in provincial spending is the result of the aid surge that took place from 2005. The analysis does not cover the period from 2009 because no subsequent poverty data are available. Further, as a large part of spending in Afghanistan is security related, we expect the provinces with higher conflict levels to have higher per capita spending (figure 2.12). Figure 2.12 Provincial relationship between conflict and spending Source: World Bank staff calculations,2011. 39 The indicator of provincial per capita spending is based on the donors’ and government’s planned/committed provincial spending in 2007–08.12 For comparative analysis, Afghanistan’s 34 provinces are classified into three groups based on their annual per capita spending: high-spending (more than US$ 200); medium-spending (US$ 100–US$ 200); and the benchmark low-spending group (less than US$ 100 per capita). To compare their development, we look at four indicators of well-being that are comparable between 2005 and 2008: food poverty, per capita food consumption, school enrollment (ages 6–12), and full immunization among children 12–23 months. Higher provincial spending appears to have only a modest relationship with development outcomes, even after controlling for conflict.13 On average, higher per capita provincial spending is associated with lower food poverty, and higher per capita food consumption and child immunization (figure 2.13). But only the relationship between per capita spending and change in food poverty is statistically significant, at 10 percent; that between provincial spending and change in enrollment (among children ages 6–12) is statistically insignificant and slightly negative. Figure 2.13 Provincial spending and development outcomes, 2005 and 2007–08 Source: NRVA 2005 and NRVA 2007–08. 12 Matt Waldman (2008): Aid Effectiveness in Afghanistan, Appendix 2 (based on the Ministry of Finance data). 13 We use “number of incidents causing civilian casualties” as the primary measure of conflict. It is based on the security incident data of the Security Information and Operation Centre (SIOC) of the UN Department of Safety and Security. The measure is constructed by aggregating (at the district level) all violent incidents in 2007–08 that involved any civilian casualty. This sum of incidents is divided by district populations to reach a more comparable measure of conflict across districts. For provincial analysis, a simple mean of the district incidents is calculated for each province. 40 More systematic analysis of the data also points toward only a modest impact of higher spending on development outcomes. To further understand this issue, the difference-in-difference approach is used to calculate the change in development outcomes. High-spending provinces have witnessed a 15 percentage point net decline (compared with low-spending provinces) in food poverty between 2005 and 2008 (table 2.5).14 But the gain in per capita food consumption is modest for high-spending provinces—a statistically insignificant increase of only Af 71 a month (equivalent to about US$ 1.40) associated with additional per capita spending of at least US$ 200. Moreover, low-spending provinces have outperformed high-spending counterparts in school enrollment. The weak impact of provincial spending on development outcomes is further supported by the results for the medium-spending provinces, which appear to have performed poorly on all indicators relative to their low-spending counterparts. Table 2.5 Change in key outcomes relative to low-spending provinces, 2005–08 Food poverty Per capita food consumption Full immunization Enrollment Provincial per capita spending (percent) (Af per month) (percent) (percent) High (> US$ 200) –16 71 2 –16 Medium (US$ 100–200) 17 –133 –4 –10 Source: NRVA 2005 and NRVA 2007–08. Returns from higher provincial spending appear to improve only marginally when the effects of conflict are taken into account. Food poverty in high-spending provinces appears to drop by another 4 percentage points compared with low-spending provinces.15 Also, the gains in monthly per capita food consumption increase to Af 110 (around US$ 2.20), still a modest gain compared with the additional per capita public spending (table 2.6). Similarly, the association between higher spending and development outcomes for medium-spending provinces shows some improvement, but the results remain worse than those for their low-spending counterparts. The evidence of incremental improvement (after controlling for conflict) is in line with the general view that conflict hampers development efforts. Table 2.6 Change in key outcomes after controlling for conflict, 2005–08 Food poverty Per capita food consumption Full immunization Enrollment Provincial per capita spending (percent) (Af per month) (percent) (percent) High (> US$ 200) –20 110 3 –12 Medium (US$ 100–200) 15 –116 –4 –8 Source: NRVA 2005 and NRVA 2007–08. The benefits of high spending appear to have accrued more to those who are better off: high-spending provinces saw an average increase of Af 70 in their per capita food consumption in 2005–08 relative to the low-spending group. This average gain can be further disaggregated by looking at the change in food consumption that accrued to the poor and better-off provinces (in terms of their per capita food consumption ranking; table 2.7). Better-off provinces witnessed a much larger increase (173 Afs) in their per capita consumption compared with their poorer counterparts (only 50 Afs). The difference in the gains of well-off and the poorer provinces shrinks, however, once the impact of conflict is taken out. 14 These calculations are based on the averages of the provincial means of each indicator, in effect treating the province as the unit of analysis. Similar patterns emerge from the estimates obtained by calculating the means across the aid categories only (without first calculating provincial means). 15 To net out the impact of conflict, the outcome indicators for 2005 and 2008 are regressed against the conflict indicator — incidents causing civilian casualties—and then the “difference in difference analysis” is carried out on the residual values of the outcome indicators. (It is assumed that the conflict levels across provinces remained broadly similar between 2005 and 2008.) 41 Table 2.7 Disaggregated impacts of high spending on food consumption groups, 2005–08 (Af per month) Poor Middle Better-off th th th Increase in per capita food consumption (25 percentile) (50 percentile) (75 percentile) Without controlling for conflict 50 50 173 Controlling for conflict 138 134 163 Source: NRVA 2005 and NRVA 2007–08. Thus a positive (but modest) association between provincial spending and development outcomes contributed to a slight increase in inequality among provinces. Given this—more so at higher levels of spending—one would expect the differential spending across provinces to add to the divergence in development outcomes over time. To explore this possibility, the inequality of food consumption among provinces is decomposed into two subcomponents: inequality of food consumption within each spending group, and inequality of food consumption between the spending groups. On this analysis,16 the share of inequality explained by the difference in per capita spending levels rises from 5 percent in 2005 to 11 percent in 2008. In conclusion, the impact of higher provincial spending on development performance appears to be relatively modest. This is particularly true when differences in per capita provincial spending are not very large. A key reason for the modest impact is the nature of spending, which is tilted toward security; once provincial conflict is considered, the results show only modest improvement. The analysis also shows that the gains associated with higher spending appear to have accrued more to well-off provinces. Some evidence also suggests that differential per capita spending has contributed to inequality of well-being among provinces. Conclusions and recommendations The picture that emerges from this analysis is cautiously optimistic. Much of the growth performance of the past decade has been aid driven. Economic growth is therefore expected to wind down as aid declines, and it will do so quite dramatically, nearly halving the average annual growth achieved over the last decade. Nevertheless, even at these rates, Afghanistan’s economy will continue to grow respectably, mainly thanks to the emerging mining activities and the continued expansion of agricultural production. And while the large aid inflows might have helped reconstruction, they have also introduced distortions in aggregate demand, as well as in labor and asset markets. Thus a decline to a more “normal” level of aid will present not only risks but also opportunities for Afghanistan to transition to a more stable, self-reliant, and sustainable economy. But with rapid population growth, GDP growth of around 6 percent implies very slow progress in raising average per capita incomes and reducing poverty: it would take around a generation to double per capita income. Current drivers of growth—agriculture and mining—will be insufficient to provide income for the increasing number of (young) people joining the labor force and to lift the large numbers of poor people out of poverty. Afghanistan therefore needs a development strategy that focuses on removing binding constraints to expansion and generates inclusive growth that promotes greater integration across economic sectors. 16 These calculations are based on the inequality in provincial food consumption. Similar pattern emerge when we calculate the inequality of food consumption across all households, grouped by spending category only. 42 The government and its development partners can do much to manage the risks to the economic outlook and support a successful transition. Donors need to reduce future aid flows gradually to avoid major disruptions. Several risks arise from a faster withdrawal of aid, including a rapid depreciation of the Afghani, high fiscal imbalances and associated adjustment costs (see chapter 3), and potentially destabilizing effects on labor markets (such as higher underemployment, lower wages, and loss of household income opportunities). The private economy and the government need time to find substitutes for official aid flows. It is therefore important that any reduction in aid is gradual enough to allow a private response to set in and, in time, fully compensate for the lower aid with investment and enhanced domestic revenue generation. In particular, aid programming should be sensitive to the time required to develop mining. It also needs to be sensitive to the type of aid that is being phased out. Given that non-salary security aid has much lower domestic content than civilian aid, it will be preferable to scale back security rather than civilian aid—provided that step does not interfere with the provision of services. Improved aid effectiveness implemented by the government and donors could offset part of the negative impact of an overall decline in aid. Increasing local domestic content can help stimulate private activity by raising demand for local goods and services. One example is “Afghan First.” Spearheaded by the U.S. government, it attempts to increase the share of Afghan vendors in aid contracting, through simplifying procurement and translating procurement documents. Shifting more aid on budget is another, even more effective, approach to increasing local content. Although its additional impact on growth might be relatively small, it is still much higher than trying to increase the local content share of off-budget spending, which is currently less than a third of on-budget spending. The government can encourage such trends by further strengthening and improving public financial management and the budget process, enforcing sound procurement procedures, and making stronger efforts to control corruption. Development partners can enhance the positive impact of aid by decreasing aid inflows’ volatility, which has been wide and unpredictable. Such swings relate not only to overall aid but also to the difference between pledges and delivery. Aid volatility may reduce the effectiveness of aid because it constrains policymakers’ ability to undertake long-term planning, and it may lead to suboptimal public spending and investment decisions (Desai and Kharas 2010). Predictability will become even more relevant for aid effectiveness the more that aid is spent on budget. Donors and the government have possibilities to manage volatility: the former by committing to longer term assistance strategies, the latter by securing monetary policy space through maintaining sufficient international reserves.17 Targeted interventions by donors and the government can support economic adjustment. The above analysis has shown that transition will result in loss of jobs. In transition-affected sectors, companies may struggle to cope with a decline in demand since the very shallow financial sector limits scope for flexible adjustments by the formal private sector. Adjustment will be made even harder by the likelihood of changing demand structures in the future. For instance, construction companies that may have specialized in building roads or providing works for military purposes may not have the required skills for emerging sectors in mining and other infrastructure. Private sector development and job creation are both long-term endeavors. Mitigating the short-term impact of the aid decline may therefore require extra measures to lessen the negative impact of adjustment processes in the private sector and labor markets. Such measures could include the 17 Eifert and Gelb (2005) argue that relatively low amounts of accumulated reserves (perhaps two to four months of imports) can smooth expenditures quite effectively under a range of levels of aid instability. 43 temporary scaling up of workfare programs or training to prepare workers for new demands arising from emerging economic activities. Efforts to sustain or improve security—as well as to bolster the legitimacy of the state—will remain of primary importance. This is because serious threats to growth and economic stability during the transition are likely to come less from declining aid itself and more from noneconomic factors. These include uncertainty over Afghanistan’s political future, falling business confidence, and the weak governance environment. All efforts of the government and its development partners to stabilize economic performance over the next few years are contingent on a stable and improving security and governance environment. 44 References Belasco, Amy (2011): The Cost of Iraq, Afghanistan, and Other Global War on Terror operations since 9/11. Congressional Research Service, Washington DC. Benn Eifert and Alan Gelb (2005): Coping with Aid Volatility. Finance & Development: September 2005, Volume 42, Number 3. Washington, DC. Peace Dividend Trust. (2011). Job Creation in Afghanistan: Putting Aid to Work. Washington, DC. US-CENTCOM. July 2008–January 2011. “Contractor support of U.S. operations in the US-CENTCOM area of responsibility, Iraq and Afghanistan.” Washington, DC. 45 Chapter 3 Managing the Fiscal Challenge Summary Afghanistan will stay heavily dependent on donor financing over the coming decade. Even if it meets ambitious domestic revenue targets—a projected rise from 10 percent of GDP to more than 17.5 percent a decade from now—the fiscal gap in the absence of continuing substantial aid would be unmanageable at 25 percent of GDP by 2021/22, though smaller than a few years earlier. The gap will stem largely from escalating security costs that threaten to crowd out development spending. Security spending, including operation and maintenance (O&M) and the salaries of 352,000 troops, will account for 17.5 percent of GDP by 2021. Without continued substantial donor assistance, these costs will be unaffordable. Introduction This chapter assesses the medium- to longer term impacts of declining aid and military spending on fiscal management. It suggests options to the government and the international donor community to manage and mitigate the adverse impacts of transition on the country’s fiscal position over the coming decade. In 2010/11, aid (including funding to the security sector) accounted for over 104 percent of GDP (figure 3.1), and only 12 percent of it was delivered on budget.18 Historically, around two-thirds of donor outlays bypass the budget. Of the on-budget third, most go toward development projects (including the national priority programs) and public sector salaries. In 2010/11, 72 percent of the operating budget was covered by domestic revenue, and the rest financed by donor trust funds. 18 On-budget aid in this fiscal analysis is defined as the total of operating and development budget grants. The figure of external budget aid at 92 percent of GDP represents donor reports to the World Bank. 46 Figure 3.1 Aid trends, Afghanistan (US$ billion) Source: World Bank staff calculations. The extremely high current level of aid (estimated at US$ 15.7 billion in 2010/11) is roughly the same dollar amount as Afghanistan’s GDP and cannot be sustained by the US and other donors. Aid has funded essential services including education and health, infrastructure investment, and government administration. Afghans’ lives have greatly improved over the last 10 years as a result. But these inflows—most of them off budget—have been so high that they have led to aid dependency, and have increased opportunities for mismanagement, waste, and corruption. Parallel systems to circumvent limited government absorptive capacity have impeded aid delivery and the building of an effective state (see chapter 1). Further, public spending financed by such high aid flows—on and off budget—will be fiscally unsustainable for Afghanistan once donor funds decline. Lesser amounts, matched by more effective aid delivery, could lead to more positive outcomes. The key issue is how to manage this change, mitigate its adverse impacts, and put aid and spending on a more sustainable path for the longer term. International experience and Afghanistan’s history after the 1989 Soviet military withdrawal demonstrate that violent fluctuations in aid, especially abrupt aid cutoffs, can be very damaging and destabilizing. Afghanistan’s heavy reliance on external funding after the Taliban’s collapse in 2001 allowed it to make considerable progress in establishing basic service delivery and a functioning government. Also, its budgetary performance in many respects has been encouraging. Sound public financial management systems have been introduced and progressively strengthened, and fiscal discipline has been maintained, including control over the civilian wage bill and staffing. But like many South Asian countries, Afghanistan’s public expenditure needs far exceed public revenue. The need to invest in physical infrastructure and enhance human capital—as well as achieve the Millennium Development Goals—is overwhelming and will become even more pressing after 2014 than in the last decade. To understand the transition’s full impact on public finances, it is important to understand the underlying factors that have driven the country’s finances over that decade. Fiscal trends since 2001 Since the fall of the Taliban, Afghan public finances have been characterized by initially low but rapidly increasing revenue mobilization, high aid dependency, and very heavy security spending. Still, public finances have come a long way. The country has made commendable progress in mobilizing domestic 47 resources and establishing a sound tax collection and administration system, and the donor-funded development budget, which has financed key national programs, expanded the rural population’s access to basic infrastructure and core services Revenue Domestic revenue collection has performed strongly, growing by over 20 percent annually in the four years to 2010/11. That year, collection reached an all-time high of US$ 1.7 billion (11 percent of GDP), exceeding the IMF target of 9.2 percent. This is remarkable considering that revenue collection stood at 3 percent of GDP before 2002/03. Currently, half the revenue collection comes from taxes, just over a third from customs duties, and the rest from nontax revenue. Taxes and customs duties have been driving domestic revenue. In 2010/11 tax collection (fixed, income, property, sales, and excises) accounted for 48 percent of total domestic revenues, growing by 29 percent over the previous year (figure 3.2). Sales tax was the largest component, reaching an estimated US$ 347 million; income tax (mostly on wages) generated around US$ 219 million; and fixed taxes (imports by licensed businesses) generated US$ 193 million. Customs duties, which accounted for 36 percent of revenue collection in 2010/11, grew by 27 percent and more than half came from duties on motor vehicles and parts and fuel imports (figure 3.3). Nontax revenues contributed 14 percent and other revenues 2 percent for the same year. Figure 3.2 Breakdown of domestic revenues, Figure 3.3 Breakdown of customs duties, 2010/11 2010/11 (percent) (percent) Other 2% Other 26% Motor vehicles and parts 28% Customs 36% Beverages and vinegar 2% Tax revenues 48% Machinery and mechnical appliances 2% Salt, sulfur, and stone 3% Iron and Minerals, fuels, steel and oils 5% 24% Electric machinery and Milling products, Nontax equipment malt starch, and 14% 5% wheat gluten 5% Source: Ministry of Finance. Afghanistan has seen a significant increase in aid since 2005. In 2010/11, aid and international spending on the Afghan National Security Forces (ANSF) was US$ 15.7 billion19—about the same as GDP. While 19 This figure excludes spending on international military forces. In fiscal 2011, for example, U.S. spending on its Afghanistan mission was estimated at US$ 118.6 billion. 48 the bulk is security spending, civilian aid is estimated at more than US$ 6 billion a year, or nearly 40 percent of GDP. Donors deliver most aid directly outside the government’s budget. In 2010/11, US$ 13.8 billion or 88 percent was executed by donors and their implementing partners outside the Afghan government budget, through the “external budget,” and only US$ 1.9 billion or 12 percent was on budget, through the “core budget.” The operating budget consists largely of the government wage bill plus non wage O&M costs, funded by domestic revenue, multi-donor trust funds, and donors’ budget support. The core development budget consists of donor-financed projects and programs executed by the government, plus a small portion funded by budget support and domestic revenue.20 Negligible 10 years ago, the core development budget now executes some US$ 1 billion of development activities annually. Expenditure The Afghan budget is heavily dependent on external financing. In 2010/11, only US$ 3.3 billion of the roughly US$ 16.9 billion in total public spending was channeled through the core budget (figure 3.4). In 2010/11, domestic revenue represented 72 percent of the operating budget and 52 percent of the total core budget, and the remainder was covered by donor grants. The development budget alone was 85 percent donor funded. Thus, even though domestic resource mobilization has made great strides, future spending needs will outstrip domestic revenue generation. Figure 3.4 Public spending—external and core budget, 2010/11 (US$ billion) Donor grants development budget 41% Other grants, CSTC-A, Operating and ADB $2.3 billion 14% 7% LOTFA 11% External budget $13.6 billion Development 80% 6% $0.9 billion 6% ARTF RCW 35% Domestic revenues Source: Ministry of Finance and World Bank staff calculations. Note: Figures represent actual development budget disbursements. The external budget estimate at US$ 13.6 billion is based on disbursement figures reported to the World Bank by donors excluding on-budget contributions to the Afghanistan Reconstruction Trust Fund (ARTF), Combined Security Transition Command (CSTCA), and Law and Order Trust Fund for Afghanistan (LOTFA). 20 In 2010/11, only 15 percent of the core development budget was domestically financed. The external budget is not fully coordinated with the government’s expenditure priorities in planning, budgeting, implementation, and monitoring and evaluation. Most donors consult with the government on off-budget projects, but the government struggles to align off-budget programs with national objectives. 49 Wages and salaries account for some three-quarters of the operating budget (figure 3.5). In 2010/11, the security wage bill alone represented about two-thirds of operations and 7 percent of GDP. In 2010, the government’s contribution to security salaries stood at 3 percent of GDP. Donors contributed as much as 23 percent of the non security wage bill.21 After salaries, the second-largest operational expense was O&M, which accounted for 19 percent of the operating budget and 2 percent of GDP in 2010/11. Finally, a further 2 percent of GDP was a combination of smaller recurrent expenses, like pensions, transfers, capital spending, and interest payments.22 Figure 3.5 Core budget development and operating expenditure (percentage of GDP) 30 25 Development 20 Percentage of GDP Others 15 O&M Civil service wages 10 5 Security wages 0 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 Source: Ministry of Finance. Security spending—mostly salaries—dominates the Afghan core budget (figure 3.6). In 2010/11, nearly 40 percent of core on-budget spending (operations and development) was allocated to security. This share will grow over the coming year as the ANSF reaches its target level of 352,000 troops by November 2012. With such high security spending and a likely decline in external financing, there is a real risk that priority development, O&M, and service delivery expenditure will be increasingly crowded out. About a third of the 2010/11 budget was allocated to education (15 percent), infrastructure (14 percent), and health (4 percent). Agriculture and rural livelihoods received 9 percent of the core budget. 21 Security wage bill contributions are paid into two trust funds, CSTCA at roughly US$ 287 million and LOTFA at US$ 520 million. Source: Fiscal Bulletin Q4 1389, MOF. The nonsecurity wage bill is paid through the recurrent window of the ARTF. 22 The total core budget for 2010/11 was US$ 4.78 billion (Af 224.781 billion), with an operating budget of US$ 3.20 billion (Af 150.726 billion) and a development budget of US$ 1.58 billion (Af 74.055 billion). 50 Figure 3.6 Core budget by sectors, 2011/12 (percent) Economic governance and private sector 4% Contingency 8% Social protection 1% Agriculture and rural Security development 39% 9% Health 4% Education 15% Governance 6% Infrastructure 14% Source: Ministry of Finance. Note: Sectoral definitions according to the Afghan National Development Strategy. Afghanistan will continue to depend heavily on donor financing and has few options to create fiscal space for new programs. With a growing fiscal gap, attaining fiscal sustainability23 will be an increasing challenge. Other countries may be in position to issue bonds or access financial markets in order to close their financing gap over time, but the Afghan government, with few financial instruments at its disposal, has limited options and scope to create more fiscal space. Mobilizing more tax revenue than envisaged (through tax policy change or tax administration reform, or by raising nontax revenue) would most likely yield marginal returns because existing revenue targets are already highly ambitious. There may be opportunities to further prioritize and reduce expenditure (mostly in the operating budget), or to improve the efficiency of public spending (or both). Operating expenses are growing rapidly, and tighter control along with efficiency gains during the transition could free domestic resources by, for instance, ensuring that procurement contracts have a higher share of local content and rely more on domestic rather than imported materials. A longer term endeavor would be reforms in private sector participation, which could lessen the burden on public investment by establishing public–private partnerships. Concessional lending is another possible avenue to raise finance, and may be a way to increase investment in infrastructure. However, the country remains at high risk of debt distress, even after the recent debt-relief efforts under the Heavily Indebted Poor Countries (HIPC) completion point. The joint IMF–World Bank Debt Sustainability Analysis in 2011 found that debt-burden indicators would 23 Fiscal sustainability here refers to the ability of government to meet its operating expenditure from domestic revenue. 51 deteriorate rapidly if the country’s large financing needs were met with loans, even on concessional terms. After debt relief under the enhanced HIPC Initiative and the Multilateral Debt Relief Initiative, the external debt burden was lightened substantially and external public and publicly guaranteed debt was US$ 1.17 billion, or 8.1 percent of GDP, in March 2010.24 Still, during the transition years, the government will need to safeguard investment to sustain real GDP growth. There is every indication that overall aid will drop in the next 10 years, and the government will have to work hard to ensure that aid tails off gradually rather than rapidly. Any rapid falloff will have serious consequences for the government’s finances (chapter 2). Medium-term projections How will Afghanistan’s fiscal position evolve over the next 10 years? This section projects the evolution of the country’s fiscal situation beyond 2014. Although the Afghan budget will enjoy greater inflows of domestic resources from mining and taxes, it will also face even greater liabilities from a growing wage bill and the O&M costs of the assets built over the last decade. Revenue Three main factors are likely to drive domestic revenue collection in 2011/12–2021/22: greater inflows from mining revenue and royalties, a value-added tax (VAT), and less reliance on customs duties. If efforts to improve tax administration continue, other tax revenue will also increase. By 2021 Afghanistan should be able to mobilize domestic resources equivalent to around 17.5 percent of GDP. This share is higher than current levels in comparable South Asian countries such as Bangladesh or Nepal. Mining will likely contribute around 2–3 percent of GDP annually in revenue and royalties by 2021/22. That contribution will increase when the two largest mining projects, Aynak and Hajigak, move from the initial construction stage (2011–15) to the operating phase (2016–22).25 Conservative estimates suggest that inflows would be around US$ 322 million a year during the first stage (US$ 205 million from Aynak and US$ 117 million from Hajigak). After 2016, when the mining projects also start to amortize their debt, revenue and royalties would be expected to increase to around US$ 923 million annually (US$ 541 million from Aynak and US$ 382 million from Hajigak).26 These estimates take into account the signing bonus, royalties, and corporate income tax on profits that the mines will generate directly. They do not consider the second-round effects of the projects becoming drivers of local output. For both mines, royalties dominate the fiscal contribution, followed by corporate income taxes and personal income taxes. 24 In present value terms, it stood at about 4 percent of GDP, or 70 percent of exports of goods and services. Public external debt service was equivalent to 1.8 percent of exports of goods and services. Debt burden indicators could deteriorate rapidly if substantial financing needs are met with loans, even concessional. Stress testing and a country-specific alternative scenario illustrate the vulnerability of the country’s economy, especially to shocks such as shortfalls in exports and protracted insecurity that hampers investment and growth (World Bank and International Monetary Fund, 2010). 25 During the construction phase, it is assumed that the mining companies pay no significant taxes except the bonus payments; during the operating phase. The profits and profit-based income taxes from both mines were based on output and price estimates that determine gross revenues. 26 This analysis assumes that the bulk of the mining resources are available for other government spending, although some of them will need to be reinvested in the mining sector and related facilities to continue driving mining growth. 52 VAT will mainly drive tax revenue increases, and public finances could receive an additional 1–2 percent of GDP by 2021. This new tax is part of the recently agreed IMF-supported program and would be brought in over three years. VAT inflows are assumed to come on budget from 2016/17. (VAT collection estimates are based on IMF projections and allow an extra year for it to come into effect.) Future tax revenue may not be as vulnerable to economic downturn from declining aid because the telecom sector, largely driven by domestic demand, is one of the largest tax contributors (figure 3.7). Telecom tax revenue accounts for close to half of large taxpayer office collection and these resources are likely to be safeguarded if demand eventually contracts. This will not be the case for construction, airlines, and services, which together account for a fourth of the tax collection and will be much more exposed to any future contraction. Figure 3.7 Breakdown of revenue from large taxpayer offices, 2010/11 Others 3% State-owned enterprise Manufacturing 14% 0.2% Telecom 45% Service 16% Hospitality 1% Airline 4% Banks Construction 11% 6% Source: Ministry of Finance, Fiscal Bulletin. Customs duties’ contribution is likely to decline from 3.8 percent to 1.3 percent of GDP by 2021/22.27 While customs is currently one of the main drivers of domestic collection, the contribution is likely to decline, reflecting a contraction in demand for transport services and trade after the transition. Annual growth in customs duties is seen falling from the current 23 percent to 3 percent by 2021/22, though the effective collection rate is assumed to continue improving slightly—primarily due to better border controls, imports documentation, and collection efficiency. These estimates are based on the assumption that security does not deteriorate further at major border posts, border controls remain stable, and further tariff increases are not introduced. 27 Customs duties are estimated to be Af 33 billion by 2021/22 assuming a gradual reduction of aid and continued progress in customs reforms. 53 Expenditure Even if ambitious targets for robust growth in domestic revenue are met (with a projected rise from 10 percent of GDP to more than 16 percent of GDP a decade from now) the fiscal gap will be an unmanageable, mainly because of higher security spending, O&M, and the wage bill, which together will account for 17.5 percent of GDP in 2021 (figure 3.8). Figure 3.8 Projected expenditure by 2021/22 (percentage of GDP) Total Afghan core budget expenditures (43 percent of GDP) Development Operating (10 percent of GDP) (33 percent of GDP) Wages and salaries O&M Other spending (16 percent of GDP) (14.5 percent of GDP) (2.5 percent of GDP) Security Security (10.5 percent of (7 percent of GDP) GDP) Nonsecurity Nonsecurity (9 percent of GDP) (4 percent of GDP) Source: World Bank staff calculations. Note: Other spending includes pensions, transfers, interest payments, and liabilities associated with Kabul Bank. The largest component of operating expenditure will be security related. Overall security spending, combined O&M costs, and an increasing wage bill will account for 17.5 percent of GDP by 2021, averaging around US$ 5.4 billion after 2015/16 in 2011/12 prices (figure 3.9). 54 Figure 3.9 Projected expenditure by 2021/22 (2011/12 US$ million) (In 2011/12 US$ 16,000 millions 14,000 Development 12,000 10,000 Civilian Operating 8,000 6,000 4,000 Security 2,000 0 Source: World Bank staff calculations. Security wages and salaries will increase from 7 percent to 10 percent of GDP by 2015/16, but will then decline to 7 percent of GDP by 2021/22, if troop numbers remain constant. The security wage bill is projected to increase by end-2012 when Afghanistan reaches its stated goal of 352,000 Afghan National Security Forces (Afghan National Army and Police)—21 percent higher than in 2010/11. The O&M needs of the security sector alone will grow to 16 percent of GDP by 2015/16 (figure 3.10). Security will require around US$ 3.5 billion a year (as estimated by NATO and excluding salaries) to cover O&M for infrastructure and equipment, existing and under construction. This figure takes into account the requirements to operate and maintain the reported US$ 11.4 billion investment by the International Security Assistance Force in security infrastructure from 2006–2014 and would cover utility bills, minor repairs, and everyday operations of the military bases and police stations. It also includes the costs to the ANSF of existing and planned fuel, spare parts, minor repairs, contractor services, and travel allowances. 55 Figure 3.10 Core budget projections to 2021/22 (percentage of GDP) 60% 50% Total, 43% 40% Development, 9.7% Percentage of GDP Other spending, 2.5% 30% O&M civilian, 3.9% Civil service wages, 8.5% 20% Domestic revenues, 17.5% O&M security, 10.6% 10% Security wages 7.4% 0% 2005/06 2007/08 2009/10 2011/12 2013/14 2015/16 2017/18 2019/20 2021/22 Source: Ministry of Finance and World Bank staff calculations. Note: Other spending includes pensions, transfers, interest payments, and liabilities associated with Kabul Bank. During the transition, the ANSF will receive a generous package of new military equipment (valued by NATO at US$ 1.4 billion) mainly comprising armored vehicles (22,000) and some 44 jets and helicopters with high operating costs. To put these costs into perspective, the fuel costs for a one-hour flight in a military jet stands close to US$ 7,000.28 So a considerable component of the US$ 3.5 billion per year will be related to fuel. Non-security spending includes the civilian operating and development budgets. After 2015/16, these budgets will average around US$ 4 billion annually, or an additional US$ 2.5 billion a year respectively at today’s prices. Non-security salaries will increase from 5 percent to 9 percent of GDP in 2021/22. About half the increase will be the combined effect of the pay and grading reform and the hiring of 10,000 new teachers annually (to reduce the very high student–teacher ratios). The reform will raise salaries for all re-graded civil servants by 2015/16. The other main driver will be the impact of bringing on budget the costs of enhanced salaries and allowances for a new senior civil service cadre in key spending ministries. By moving these costs on budget the government will attempt to make the civil service less vulnerable to the vagaries of external funding. Line ministries will be required to increase spending on O&M to maintain existing and new assets that they will become liable for during the transition. These O&M expenses, or code 22 in Afghanistan Financial Management Information System, represent the costs required to maintain assets at their current performance, while excluding wages and salaries and consulting costs. They include: minor repairs, contractor services, travel allowances, and utilities, fuel, spare parts for equipment. In 2010/11, the share of the budget dedicated to O&M liabilities (both from the security sector and civilian assets) stood at only 2 percent of GDP. 28 NATO Training Mission Afghanistan, Ambassador’s Conference, May 2011. 56 The O&M costs of civilian assets will increase from 1 percent to 4 percent of GDP by 2021/22. The recurrent liabilities of maintaining civilian assets are around US$ 1.3 billion a year, excluding salaries and major rehabilitation expenses (Altai Consulting 2011). These estimates, which are based on examining seven key asset categories, are conservative and dependent on three variables: the expected level of service from the asset, the assets’ condition when handed over to the government, and development spending over the coming three years. Based on these factors, one can assume a minimum level required to sustain the current service delivery, but without factoring in rehabilitation investments.29 Other smaller operating liabilities will represent an additional 2 percent of GDP by 2021/22. These include pensions, capital expenditure, interest payments, and transfers and subsidies. Within this component are the liabilities of the bailout and refinancing of the Kabul Bank, which represent around US$ 900 million to be paid over the next eight years and serviced at an interest rate of 2 percent annually. The small capital investment component is estimated to accrue on average 0.6 percent of GDP. The Afghan government has limited access to external financing, so this analysis assumes that the budget impact from interest payments will remain marginal over the transition. This could change in the future if the country develops a sovereign bond market. The liabilities from pensions associated with the increasing security and civilian wage bill do not factor into this fiscal analysis because they will require a much longer time to have a noticeable impact (30–40 years), and will depend on the composition of the ANSF and civil service at that time. The development budget is projected to rise from 6.3 percent to 9 percent of GDP over the period. Development spending will most likely continue to grow and donors will increase their on-budget contributions, fulfilling their Kabul Conference commitment of shifting more of their aid on budget, while reducing off-budget aid. But absorptive capacity and budget execution issues will likely keep the increase gradual. Thus these projections assume a moderate 14 percent annual rise up to 2021/22. Closing the financing gap How will these fiscal developments shape fiscal management over the coming decade? Given that domestic revenue will be insufficient to cover spending and aid will likely decline, how can Afghanistan close this financing gap? The total core balance is projected to reach an enormous financing gap of 33 percent of GDP by 2014/15 and ease to 25 percent of GDP by 2021/22 (after the bulk of the mining revenue materializes) (figure 3.11). The operating budget gap alone would be 16 percent of GDP by the end of the period. To put these figures in context, the financing gap that triggered Greece’s current fiscal crisis was close to 13 percent of GDP. Clearly, Afghanistan’s projected fiscal gap is unsustainable—and in any case could not be financed in the absence of sizable aid. 29 The calculation of civilian O&M needs is based on central and provincial research commissioned by the World Bank for this report. The research required the construction of a comprehensive public asset registry to determine the bulk of the existing and soon-to-be handed over assets in need of O&M over the coming years. Fieldwork was conducted in the provinces of Uruzgan and Bamyan. (For more information on O&M needs and estimation, see Altai Consulting 2011.) 57 Figure 3.11 Operating and total core financing gap (percentage of GDP) 10% Projection 0% -5% Operating balance with security Total core balance without -11% -10% security Percentage of GDP -16% -20% -25% -30% Total core balance with security -40% -50% 2002/03 2004/05 2006/07 2008/09 2010/11 2012/13 2014/15 2016/17 2018/19 2020/21 Source: Ministry of Finance and World Bank staff calculations. Note: Projections assume that domestic revenue equivalent to 3 percent of GDP is devoted to security costs, and the rest to civilian expenditure. If the security sector (Afghan National Army and Police) remains at currently agreed levels, Afghanistan will have to rely on continued international funding to pay most security spending (salary and nonsalary). The share of aid channeled through the government budget will need to rise sharply. Indeed, virtually all aid projected in 2021/22 (equivalent to 25 percent of GDP or US$ 7.2 billion, and averaging US$ 7.8 billion for the intervening years in 2011/12 prices) would need to be on budget (figure 3.12). Thus, strict priorities will be required to contain nonsecurity budget expenditure within a fiscally manageable resource envelope—civilian aid, for example, will need to be spent much more selectively for development and O&M. Even so, Afghanistan’s fiscal situation will remain fragile and vulnerable to external shocks. 58 Figure 3.12 Total core financing gap (in 2011/12 US$ million) 0 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 -1,000 -2,000 -$2,500 -3,000 -$3,200 2011/12 $ million -4,000 -5,000 -6,000 -$5,700 -7,000 -$7,300 -$7,500 -8,000 -$7,800 -$7,700 -$7,700 -$7,900 -$7,900 -$8,400 -9,000 Source: World Bank staff calculations. It is possible to envisage various combinations of aid and domestic resources to cover the spending gap. A reasonable approach could be for domestic revenue to cover the full civilian operating budget —wages and O&M—and part of the security costs (at the 2010/11 level of 3 percent of GDP). Donors would finance the remaining bulk of security costs plus a more highly prioritized core development budget. Other low-income countries receive, on average, around 9 percent of gross national income in non- security development assistance. Afghanistan would require close to three times that in total aid to plug its projected financing gap. If achieved, it could have a balanced budget by 2021. But if just 16 percent of GDP in aid materialized, this would only finance the operating budget (see figure 3.10). Lower, more rapidly declining aid would force the government to make even more difficult trade-offs between security and civilian spending. If aid is 10 percent of GDP lower and the security sector is protected at the currently targeted size and cost, development spending would have no room at all, and civilian O&M would most likely be squeezed as well. But if the civilian budget is protected, security could only be funded at less than half of currently targeted levels. If cuts in aid are even deeper, particularly if they are in line with the rapid decline in aid scenario (with aid projected at 10 percent of GDP from 2018, see box 2.3), the trade-offs would become all the more stark and damaging. So it will be very important for the government, working closely with donors, to ensure that the core civilian budget does not become a casualty of high security costs and inadequate aid. Increasing on-budget aid and managing O&M through government systems would greatly improve aid effectiveness. But in addition to further improving PFM and governance in order to provide donors greater confidence to shift more aid on budget, the government will need to overcome serious absorptive capacity constraints if it is to be in a position to receive additional donor money on budget. After core development spending more than doubled in absolute terms between 2005/06 and 2007/08, 59 disbursement has stalled at just under US$ 1 billion annually over the past four years (figure 3.13). In contrast, operating budget execution has been high, but was almost all wage spending. Thus, low execution rates are a combination of structural and capacity issues: weak budget formulation by line ministries and implementing agencies that led to unrealistic budgets and work plans; varying capacity for implementing projects and keeping to procurement procedures; poor predictability and late disbursement of donor funds; and security concerns that make project implementation difficult. In the future, managing O&M—expected to increase from US$ 335 million now to US$ 4.8 billion by 2015/16 (including US$ 1.3 billion for civilian O&M)—will be a major challenge. Figure 3.13 Core operating and development budget execution, 2005/06–2010/11 (percent and USUS$ million) 100 96% 98% 1,200 93% 94% 96% Operating 90% 90 1,000 80 70 800 $ million 60 54% 54% Percent 50 Development 600 43% 40 41% 39% 40% 400 30 20 200 10 0 0 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 Source: Ministry of Finance. Note: Budget execution relates to the amount of the approved budget that is actually disbursed. Conclusions and recommendations The greatest adverse impact of transition will be fiscal—Afghanistan faces a projected financing gap of 25 percent of GDP by 2021/22, even higher in some of the intervening years. The fiscal gap will be enormous despite hoped-for robust growth in domestic revenue (projected to rise from 10 percent of GDP to more than 17 percent a decade from now). This gap arises primarily due to high wage and 60 nonwage security costs (projected at around 17.5 percent of GDP in 2021/22), along with funding the government and sustaining service delivery on the civilian side. The civilian wage bill is projected to increase to 9 percent of GDP, the non-security O&M expenditure to 4 percent, other operating spending to 2.5 percent, and the core development budget to 10 percent. To close this financing gap, Afghanistan must rely on continuing international funding to pay for most security costs. Even setting the size of the ANSF on a somewhat lower trajectory than previously targeted (352,000 personnel with an estimated sustaining cost of US$ 5 billion a year) will not substantially change the requirement for continuing large-scale external funding of the security sector and the need for government to make tough choices on how it spends its limited budget. Making choices in a highly constrained resource environment is not easy for any government, but it is especially difficult in the current political and security situation in Afghanistan. Even so, it will be essential that the government, working closely with donors, ensures that the core civilian budget does not become a casualty of high security expenditure and falling aid. A reasonable approach could be for domestic revenues to cover the civilian operating budget and part of the security costs (at the current level of 3 percent of GDP). Donors would finance the remaining bulk of security costs plus a more highly prioritized core development budget. Without continued and substantial international funding for security—even if security costs decline— the government will not be able to pay for its security forces and equipment, nor have any money for its development budget. In such a scenario, past development gains would not be maintained, potentially risking instability if people’s expectations went unmet. The total future package of core basic services to be maintained will depend on government and donor choices about what they can afford and their priorities going forward. And with civilian aid likely to drop well below the current US$ 6 billion annual figure, Afghanistan will be hard pressed to continue funding all its civilian programs at current levels. Yet the room for government to maneuver will likely be constrained by political, social, and economic considerations. So any chance for fiscal viability will require continued international funding for most wage and nonwage security costs over the medium term. Facing large demands for security spending, the government will also need support in protecting a more prioritized civilian budget, including public investment, basic social services, and O&M. At the same time, donors will need to channel a much larger proportion of aid flows through the government budget. This will help mitigate some of the adverse economic impacts of declining aid because the local economic impact of on-budget aid is much higher than aid channeled through the external budget. But government’s absorptive capacity is limited. Budget execution has been stuck at around a US$ 1 billion dollars annually for the past four years. Without improvements in budget execution the government’s ability to absorb additional money on budget will be severely constrained. Over the next few years a concerted push by donors and government alike is needed to improve government capacity to spend its development budget. 61 References Altai Consulting. 2011. Assessing O&M of Key Assets in Afghan Provinces. World Bank: Kabul World Bank and International Monetary Fund. 2010. “Islamic Republic of Afghanistan: Joint World Bank/IMF Debt Sustainability Analysis, 2010.” Washington, DC. 62 Chapter 4 Building Government Capacity: Key Issues and Emerging Challenges Summary Afghanistan has been relying very heavily on outsiders to deliver services. A decade on from 2001, donors continue to fund large numbers of technical assistance personnel at high cost. A survey of eight key ministries and one agency by the World Bank in 2011 found just under 5,000 externally funded staff (EFS) working in these entities, all but 200 Afghans, estimated to cost US$ 125 million a year. Capacity development initiatives have largely benefited Afghans working on donor projects (many of which are delivered through the external budget). The key to building long-term government capacity is to transition this support in a more coordinated and strategic manner to government institutions. Such transitioning will not be easy in the face of strong vested interests among Afghan and international beneficiaries—to maintain the current position, and will require government commitment to reform as well as concerted donor effort. Less aid may provide an opportunity to achieve better and more coordinated assistance in developing critical government capacity. The likely main effects of a gradual decline in aid and the shifting of aid on budget over the next few years will be a tightening job market and downward pressure on wages, less technical assistance delivered through parallel systems and more through government institutions, stronger links to public administration reform as a whole, and improved regulation and transparency of technical assistance. Too rapid a decline in aid could, however, lead to an exodus of consultants from the country, which might seriously undermine government capacity and service delivery efforts over the short to medium term. The government needs to work to retain the confidence of skilled Afghans to stay. Donors need to better align their support with government efforts to rationalize and harmonize salaries in the second civil service of Afghan EFS. Without such alignment, developing core government capacity will be difficult. Introduction By 2002, years of conflict had seriously eroded Afghanistan’s public sector. Most skilled Afghans had escaped during either the Soviet occupation in the 1980s or the mujahideen wars in the 1990s. A report in early 2002 describes the situation: “Even at the central level in Kabul, ministries or departments are war damaged shells, without even the most basic materials or equipment, and with few experienced staff. Before the Taliban 43 percent of government employees were women, most of whom were dismissed. … Government staff was paid intermittently, if at all, and many of the senior personnel have 63 either left the country or sought alternative part-time employment” (ADB, UNDP, and World Bank 2002: 15). An administrative structure still existed in many parts of the country after the war, but few senior staff were left in Kabul to manage relations with the provinces, and even fewer resources remained to deliver any services.30 Given the low skills and the need for an immediate impact, donors launched initiatives to support staff and raise government capacity. Many efforts relied more on substituting for civil service capacity than strengthening it, and often included the following modalities: regular civil servants on standard government terms; civil servants receiving higher salaries or top-ups via government or donor- funded initiatives; contract positions in the civil service filled by national or international consultants; national staff in NGOs, the United Nations, and international agencies on secondment to the government; and contractors employed either directly or through donors to carry out development projects. In the last 10 years, the international community has invested greatly in reconstructing the Afghan state, but progress in building capacity in government institutions has been slow, and much of this investment has bypassed the civil service.31 Most capacity has been built among contracted staff of donor-funded projects. A 2006 review of World Bank support to fragile states noted the following of Afghanistan: “The ‘buying of capacity’ through massive technical assistance has not delivered capacity development, and some evidence suggests that it detracted from this objective. The amount of technical assistance provided to date is well beyond the country’s absorptive capacity” (World Bank 2006: 23). Writing in 2007, Serge Michailof observed, “There is a widespread dissatisfaction in Afghanistan with the high cost of technical assistance and its limited impact in terms of capacity building” (Michailof 2007: 1). A 2009 OECD report expressed similar views.32 Since 2009 the United States in particular has greatly raised the number of its civilian advisers in the country, responding to President Obama’s 2009 Afghanistan strategy,33 which announced a new civilian as well as military surge. Despite this and other international efforts there is little indication that longer term capacity-building objectives are much nearer to being achieved.34 According to a 2010 report, “donor practices, including paying significantly higher salaries than the Afghan government can sustain and providing salary support outside Afghan planning and budget processes, distort the local labor market and undermine long-term goals of building government capacity and fiscal sustainability.”35 A recent U.S. Senate Foreign Relations Committee Report noted that U.S. overreliance on international technical advisers to build Afghan capacity could actually undermine these efforts: “Our aid projects need to focus more on sustainability so that Afghans can absorb our programs when donor funds recede” (U.S. Senate Foreign Relations Committee 2011: 3). 30 A World Bank/AREU study of subnational administration in 2004 found, surprisingly, that many of the structures and staff in the provinces were still in place despite the lack of an ongoing relationship with Kabul (World Bank 2004). The report observed that many of the staff were aging and about to retire, potentially taking much of their knowledge base with them. The opportunity to use this platform was never taken up by the government or international community. 31 According to the Afghan Ministry of Finance, since 2002 the international community has disbursed nearly US$ 57 billion in external assistance to the country, of which over 50 percent was to the security sector (Ministry of Finance 2011). 32 “A widely held view within the government is that very little donor technical assistance has been directed specifically at building government capacities for program delivery and for increased public accountability. The perception, therefore, is that ineffective technical support has contributed to the weakening of local ownership and leadership of the development agenda for the country and has inhibited efforts to build capacities within the national government” (OECD 2009: 29). 33 Announced in speeches by President Obama on March 27 and December 1, 2009. 34 GAO (2011a) states that the overall impact of U.S. efforts are unknown because of lack of performance data. 35 SIGAR Audit-11-5, 2010. 64 The next three years offer an opportunity for both the government and international community to develop a more strategic and sustainable approach to rebuilding public capacity. The transition provides an opportunity to rethink how capacity is delivered and to reduce the dependency of government on ad hoc and often fragmented technical assistance interventions. Less aid in this instance may provide an opportunity to achieve better and more coordinated assistance in developing critical government capacity. International experience Afghanistan’s experience of capacity development over the last 10 years is far from unique. The immediate post-conflict experience of most low- and middle-income countries—such as Georgia, Liberia, Serbia, Sierra Leone, and Timor-Leste—has been similar to Afghanistan’s, especially in their initial widespread dependency on mostly ad hoc, donor-driven technical assistance and salary supplementation to fill the vacuum left by the lack of government capacity immediately after the conflict. During the immediate transition local consultants’ salaries often increase sharply, as the best and brightest leave the civil service (if they had not already) to join donor-funded programs or NGOs on much higher wages. An OECD report in 2009, for example, accuses donors in Timor-Leste of “undermining the Government’s ability to perform, by engaging in ‘capacity stripping’” (OECD 2009: 50), and a 2011 OECD report talks of the inflationary impact on local wages of competition from international agencies for the limited pool of skilled national labor. Rebuilding the civil service after such interventions is therefore extremely difficult without donor-supported salary-supplement schemes, as most skilled civil servants show little interest in working directly for the government. In Liberia and Sierra Leone, donor-funded parallel policy and management systems bypassed weakened (and initially largely nonexistent) civil services (UNDP and World Bank 2011). In Liberia, the Governance and Economic Management Assistance Program formally transferred cosignatory authority to technical assistance staff in key public institutions to ensure transparency and accountability of public funds (UN– DPKO and World Bank 2006). In both countries donors had to fund salary top-up schemes to attract key individuals back from the diaspora. In Timor-Leste many advisers still retain line management positions, and technical assistance interventions have tended to focus on developing individual civil servant capacity rather than strengthening institutional capacity in the longer term. For many years donors have drawn talented Timorese away from the public services by paying higher salaries. The government has recently started to attract these staff back with enhanced salaries (OECD 2011). In an attempt to harmonize pay arrangements the Australian government has unilaterally introduced a standard remuneration framework for advisers (OECD 2011). In Georgia after the 2003 Rose Revolution the government lacked the resources to pay salaries to the skilled professionals it needed to push through reform. Through a donor-supported capacity-building fund it hired staff to support reforms (UNDP 2007). Once domestic revenue had sufficiently recovered, the government took over the salary-support scheme for top officials. In Serbia the initial response of donors after the downfall of the Milosevic regime in 2000 was to establish a fund to pool and coordinate multi-donor support for institutional development and reform. Yet rather than spearheading wider public administration reform, the fund quickly became a staffing facility to buy technical expertise in an ad hoc and nonstrategic way—the fund enabled experts to meet 65 immediate needs but did not build public long-term capacity. The fund’s failures did, however, lay the basis for a more comprehensive approach to public administration reform. In all of these examples, the challenge for both government and donors has been to wean themselves away from short-term interventions as early as possible. Bringing in consultants (foreign or local) to do the jobs of civil servants is relatively easy, and while these initiatives may be needed at the time, they are rarely sustainable in the longer term, and in the short term can undermine reform incentives. Nevertheless, countries’ ability to rapidly move away from such ad hoc, short-term interventions is sharply constrained by their particular job markets and access to resources. First, the depth of the skills deficit in each country is different, and this will largely determine the balance between dependency on foreign or diaspora capacity and use of local capacity. Second, the domestic budgets of low-income countries are more constrained. Without long-term donor support they cannot afford to pay for skilled staff. Both Georgia and Serbia could eventually pay for the higher salaries of senior staff from their own budgets, but for low-income countries like Liberia and Sierra Leone (and Afghanistan) this is much harder. Third, competition among donors drives up the price of what is already a limited pool of skilled labor, making it more difficult to either retain or attract staff to the regular civil service. Fourth, the provision of technical assistance becomes an end in itself rather than a link to wider institutional reform to create a more capable civil service. Lack of both donor coherence and government leadership undermine public administration reform efforts. Some of the key international lessons for Afghanistan are summarized in box 4.1. Box 4.1 Some key international lessons for Afghanistan  Donor capacity-building interventions need to support a common and agreed upon administrative reform plan.  Provision of external technical assistance needs to be based on clear assessment of need.  Technical assistance should have built-in sunset arrangements and be tied to specific, defined objectives (including “transfer of knowledge”).  Salary supplementation schemes need to be managed by the government (not the donors) and made fully transparent with published pay scales and allowances.  Such schemes should be part of a broader strategy to address short-term constraints and build long-term capacity.  They should be of limited duration and must be transitioned into the national wage bill as part of the normal budget process as early as possible.  Salary supplements should be linked to the function rather than the person to allow for more strategic focus.  Donors should agree to harmonize the salaries they pay to national consultants so as not to undercut government efforts to attract and retain skilled staff. Source: UNDP 2007; OECD 2009 and 2011; World Bank and UNDP 2011. The World Development Report 2011 recognizes that even the fastest institutional transformations have taken a generation. The report asks, “How much time has it taken to move from current levels in fragile states…to a threshold of ‘good enough governance’?” (World Bank 2011c: 108). In response, it notes “It took the 20 fastest moving countries an average of 17 years to get the military out of politics, 20 years to achieve functioning bureaucratic quality, and 27 years to bring corruption under reasonable control.” The prognosis for Afghanistan is not all bad. How donors and the state handle the opportunities and challenges presented by the transition are critical over the next few years. 66 Building a more effective Afghan state Technocratic approaches to state building in Afghanistan have historically had to contend with the nature of politics in the country, where formal office and position are used as resources to balance competing elite interests. The Afghan state—while having a highly centralized, unitary character as embodied in successive Constitutions—has always had weak central control and has needed to build coalitions of common interest with a strong periphery. The use of state position and office as bargaining tools in the wider political process has a long history, ensuring that attempts to introduce modern, merit-based public sector reforms face an uphill struggle (chapter 1). In the future, the importance of capturing formal office and position as a tool for accumulating wealth, and hence more political power, may intensify in the face of reduced western patronage and a shrinking war and aid economy (Surhke and Hakimi 2011). The tensions between a highly centralized de jure state and a strong de facto periphery, and between patronage and merit-based models, lie at the heart of public administration reform in Afghanistan. Progress in building at least the outward signs of a modern bureaucracy has been striking, but these reforms have failed to closely link nascent institutional developments to improvements in state capability and legitimacy. A critical aspect of building this capability in the future will be, i) to delegate greater budget authority and accountability to sub-national units of government, and ii) to develop core civil service capacity both at the centre and at sub-national levels to achieve better service delivery and budget execution outcomes in an increasingly tight fiscal environment. Sub-national administration Sub-national administration in Afghanistan is characterized by the above tensions and weaknesses in capacity. Direct formal functional and budget authority for the delivery of most key services in the provinces—such as education, health, water and roads—is held by central line ministries which work in vertically integrated silos with relatively weak linkages between them. However, this system co-exists with – and is cut across by - the provincial governor system which allows provincial governors, who are political appointees, to intervene (albeit sometimes unofficially) in the affairs of the line ministries and other agencies36. In the traditional bureaucratic structure, provincial governors were mid-level Ministry of Interior appointees who were mainly responsible for police and security matters. After 2005, however, and with the expansion of the Taliban insurgency in the south of the country, key donors started to place more emphasis on building up the authority of provincial and district administration, in particular the authority of Governors’ offices. This was driven largely by security and stabilization considerations in conflict affected areas, and recognition of the limited capacity of the center to deliver services to outlying areas. The creation of the Independent Directorate of Local Governance (IDLG) in 2007 consolidated this trend, with work beginning on a Sub-national Governance Policy in the middle of 2008. This policy was eventually endorsed by the Cabinet in 2010. In spite of this endorsement, there appears little ownership of the policy across government and continuing uncertainty about the exact roles and relationships of the different key sub-national bodies, including provincial and district governors, line departments and local councils, to each other. The policy and programmatic decisions that have sought to strengthen the link between the central Government and the provinces have created an increasingly uncoordinated and complex web of formal 36 See World Bank (2008), Intergovernment Fiscal Relations and Sub-national Expenditures in Afghanistan, August 2008. 67 government structures, locally elected bodies, donor-conceived and funded initiatives, and formal and informal local institutions of service delivery and accountability. Some of these are the responsibility of the IDLG but others, such as the NSP’s community development councils (CDCs)37, which are proposed at some future date to become Village Councils38, have been established by the Ministry of Rural Rehabilitation and Development (MRRD). This has often meant competing institutional agendas and lack of an overall strategic framework for addressing accountability and planning issues at local level. The sub-national administrative bureaucracy (especially the provincial departments of line ministries but also the Governors’ offices) remains poorly resourced and under-skilled to fulfill many of the planning, monitoring and reporting functions being suggested for it. Most significantly, there remains minimal linkage between local planning systems, budgetary resources and decision-making processes, rendering largely theoretical much of the exercise to extend the state down to the local level. Current efforts to increase provincial participation in budgeting and to adopt a more equitable basis for allocations to provincial line departments are taking place without a wider debate about what is a fiscally sustainable model for sub-national governance in Afghanistan. With some 364 districts in the country the ability of government to maintain and fully staff its districts as well as resource thousands of future village councils is heavily constrained. Currently, most financial flows to provinces and districts are outside the budget and have been heavily skewed toward conflict affected provinces which have received the bulk of donor aid flows. In 2010/11, for example, Helmand Province received over three times the resources “off budget” that it received through the government’s core budget. This is all set to change. Transition will reduce the flow of these “off budget” resources, which is likely in turn to reduce the levels of current service delivery in these provinces (see chapter 5). This may well undermine the gains to state legitimacy that have been made in recent years in some of these areas, as well as the patronage of local Governors that has been built up on the basis of direct access to donor and PRT resources. On the more positive side, however, any significant drop in “off budget” resources should help re-confirm the centrality of the government’s budget and budget planning process, and provide opportunities to develop a more equitable basis for budgetary allocations across the country in the future. The challenge for government will be to try, within a much more constrained resource environment, to ensure a more participatory and accountable budget process that does not raise unnecessary expectations, while at the same time improving the efficiency of service delivery. The regular civil service The government has contained the size of the non-security civil service: numbers have grown slowly from 327,000 in 2004 to around 355,000 in 2011. This may reflect both the government’s narrow fiscal space once it has paid for the very large security sector and its heavy reliance on non-civil servants to deliver services, rather than any overt desire to cap civil servant numbers. Still, the net effect is a relatively small civil service by regional standards.39 37 CDCs are village-level elected councils created by NSP in 2003 as local governance and service delivery institutions. By 2012, more than 70 percent of rural communities had been mobilized to establish CDCs, equivalent to nearly 25,000 councils across the country. The CDCs are responsible for implementation and supervision of development projects and liaison between the communities and government and nongovernmental organizations. 38 In the continuing absence of Independent Electoral Commission sanctioned village elections for Village Councils an inter- ministerial committee established after the Kabul Conference in June 2010 recommended CDC’s should become de facto interim Village Councils. This proposal has still to be endorsed by the Head of State. 39 Depending on the population estimate, the ratio of civil servants to population in Afghanistan is about 1.3 (based on a population of about 26 million) compared with 1.97 in Pakistan, 3.04 in Sri Lanka, 0.62 in Bangladesh, 1.23 in Nepal, and 1.25 in India (World Bank estimates). 68 The total non-security wage bill in the 1390 (2011/12) budget is US$ 765 million, less than half the security sector wage bill of US$ 1.6 billion. This non-security wage bill is set to grow from 5 percent to 8 percent of GDP from 1389 (2010/11) to 1394 (2016/17) after the completion of pay and grading (P&G) reform, further teacher recruitment, and increased incentives. Donors currently meet about 23 percent of non-security payroll costs, largely through the Afghanistan Reconstruction Trust Fund’s (ARTF) recurrent cost window. Approaching 2014 and looking beyond, the government will increasingly be expected to take on greater responsibility for funding this workforce and absorbing some contract staff currently funded by donors through the development budget, which will put further pressure on the budget. But the government’s ability to do this will be seriously constrained by the continuing growth of the security sector, which squeezes out non-security costs, and by the need to take on board O&M liabilities from the development budget. In such a tight fiscal environment, donors need to continue paying for the civil servants who are critical to service delivery up to and beyond 2014, but only as part of a broader plan to reform the civil service and rationalize the current ad hoc technical assistance and salary-support arrangements. Several civil service reform programs were carried out in the 2000s.40 The Independent Administration Reform and Civil Service Commission (IARCSC) was established in 2003 with a mandate to lead civil service reforms. The priority reform and restructuring program (PRR) was introduced to improve how critical departments in key ministries functioned and to enable these departments to recruit staff on merit, and at modestly better pay. Other activities included promulgating laws and regulations for the civil service. The government introduced a more comprehensive Civil Service Reform Program in 2007. This supported the rollout of the pay and grading (P&G) reforms for civil servants and modest restructuring of ministries and agencies. It also aimed to strengthen human resource capacities in selected ministries. The government worked with USAID to set up the Afghanistan Civil Service Support Program, which sought to boost human resource capacity, improve recruitment in provinces, and provide training. These later projects built on lessons learned through the priority reform and restructuring exercise: P&G, for example, now targets all government ministries and agencies for reform, not just a few. Civil servants are still paid in a variety of ways. These include non-reform salaries (those that have yet to go through P&G); interim additional allowances paid through priority reform and restructuring and its “superscale”; presidential awards for deputy ministers, chiefs of staff, and general directors; cadre allowances for specific jobs such as principals and senior doctors; topped-up salaries for employees of revenue-generating ministries; senior staff employed through the ARTF-funded Management Capacity Program (MCP), introduced in 2007; specific pay structures for members of parliament; and P&G salaries paid through the Civil Service Law of 2008 (Appendix 3). To date, around 303,000 positions out of a total 355,000 civil service (nonsecurity) positions have been regraded, demonstrating that reforms in ministries have to some extent been successfully rolled out, but much effort is still required to harmonize the different payment mechanisms and salary scales. Since 2008 P&G reform has been a central component of public administration reform, aiming to streamline organizational departments in ministries and their staffing structures (known as a ministry’s tashkeel). It has also been the selected instrument to migrate a closed, career-based civil service to an open, merit-based one, where all posts can, in theory, be open to competition. This allows human capacity to be injected from the outside at different levels, rather than at the entry point only. Nineteen ministries and seven agencies have completed P&G reform at central and provincial levels. It is planned 40 A comprehensive account of these programs can be found in World Bank (2008). 69 that all civil servants will receive P&G salaries by 2014, achieving two key objectives of reform: decompressing older pay scales where all civil servants were essentially paid a flat salary with allowances that were open to abuse, and completing the transition to a merit-based civil service. P&G rollout in the provinces is higher than expected, with 94 percent of staff that have completed the reform process based there. An impressive achievement, but these results are driven by the Ministry of Education, which has 105,000 teachers receiving P&G salaries across the country. This preponderance of P&G staff in the provinces is especially marked for the lower grades, 5–8 (table 4.1). There are still too many senior civil servants in Kabul, reflecting the overemphasis on building capacity there rather than at the periphery, where most Afghans live. This bias is partly because ministries’ internal incentive structures favor those close to the seat of power. On the contrary, provincial service must come to be seen as a necessary step on the career path of every senior civil servant. Effective frontline service delivery requires greater numbers of senior staff in the provinces. Table 4.1 Grade breakdown of pay and graded civil servants at the center and provinces Grades 1 2 3 4 5 6 7 8 Center 76 469 1,343 4,102 10,719 8,255 5,929 5,736 Provinces 46 1,115 1,979 7,327 24,266 24,844 13,605 16,627 Source: IARCSC data. Note: Information available as of mid-2011. P&G salaries, ranging from US$ 100 in the bottom grade to over US$ 700 in grade 1, are a huge improvement on the previous salaries, which ranged from about US$ 50 to US$ 200 a month, but they still cannot compete with non-civil service salaries. The government has struggled to retain its best staff, who quickly leave to join the burgeoning aid sector or return to civil service at much higher salaries, working as contracted-in staff paid by donors. Ashraf Ghani, Afghanistan’s former Minister of Finance, has been a long-standing critic of donors for fueling this “second civil service.” In 2009 he wrote: “Due to enormous differences in the pay of international organizations, there has been a steady brain drain from the civil service. … A fraction of resources currently used on international advisors and technical assistance could be used to recruit top civil servants” (Ghani 2009: 17). Technical assistance One needs to draw a distinction between, on the one hand, technical assistance provided by donors directly to government ministries and agencies, as well as to on-budget government projects, and, on the other, technical assistance provided to parallel institutions, like PRTs, and to off-budget donor projects. It is the drawdown of the former that will have the most immediate impact on government services, as the personnel involved often directly substitute for government capacity. Strictly, the second civil service refers to this category. Within this category are two further distinctions, between those functioning in actual line management positions as temporary tashkeel civil servants and those outside the tashkeel, who may still work in government ministries or implementing projects, but as non-tashkeel contract staff. While both may be paid by donors through a variety of salary schemes, the former are considered civil servants, the latter are not. The extraordinary costs associated with deploying EFS in Afghanistan are largely associated with the off-budget international staff that are employed in many different ways either directly by donors or through subcontractors. 70 Ten years on from 2001, donors continue to fund large numbers of external technical assistance personnel at considerable cost. Excluding the security sector,41 the OECD’s Development Assistance Committee estimated donor spending on technical cooperation at over US$ 605 million in 2008, the last year for which it has complete figures, representing 12 percent of the US$ 4.8 billion in total donor assistance disbursed that year (table 4.2).42 The figures in the table need to be treated with caution, as they show wide annual fluctuations without any obvious explanation. From 2005 to 2007, for example, the absolute value of technical assistance fell by nearly half, but total donor assistance increased by almost a third. Table 4.2 Total donor assistance and technical assistance Total donor assistance Technical assistance Percentage of total donor Year (US$ million) (US$ million) assistance 2001 404.64 — — 2002 1,300.49 255.20 20 2003 1,590.70 160.80 10 2004 2,303.10 120.30 5 2005 2,817.89 1,092.90 39 2006 2,955.78 916.30 31 2007 3,964.60 566.60 14 2008 4,865.08 605.00 12 Source: OECD-CRS data. The OECD figures are unlikely to pick up the extraordinary security costs associated with deploying civilian advisers to Afghanistan. One U.S. government civilian adviser, for example, costs US$ 425,000– US$ 570,000 to deploy to the country (SIGAR 11-17-2011). Between 2009 and 2011 the United States alone spent US$ 2 billion sending an additional 720 civilian advisers to Afghanistan as part of the civilian surge. As of March 2011, more than 3,000 Americans were employed by USAID projects in the country at a conservative annual cost of US$ 750 million.43 Total annual costs for civilian technical assistance are likely to run into several billions of dollars, if we include all advisers funded by donors working in PRTs, NGOs, contracting companies, and the government. To take one example, one PRT in a heavily conflict-affected province has some 260 staff, and half are civilians. The cost, including security and subsistence, is US$ 42 million a year just for the civilian staff of one of the several countries whose staff are located in the PRT. On top of this are ancillary costs for helicopter transport, as it is unsafe for staff to use ground transport. Including all civilian staff in the PRT the overall sustainment costs would easily reach US$ 100 million a year. This represents more than the entire government operating budget for this province (including ANSF salaries) in 2010/11. In the civil service, several programs have been launched in the last decade to build or inject capacity. The government initiated the Technical Assistance and Feasibility Studies Unit in 2003, with support 41 The largest number of technical assistance personnel in Afghanistan are the 139,000 foreign troops in the country, who are fighting an insurgency and training the ANSF. Costs are likely well in excess of US$ 100 billion a year, which dwarfs all civilian technical assistance. (Congressional Research Service 2011). 42 Technical cooperation is defined by OECD/DAC as financing of activities whose primary purpose is to augment the level of knowledge, skills, technical know-how, or productive aptitudes of the population of developing countries. It includes the cost of personnel, training and research, as well as associated equipment and administrative costs. The figures for Afghanistan are calculated from OECD-CRS data. 43 Numbers of Americans working on USAID projects taken from USAID Portfolio Review, June 2011. Costs are calculated at a conservative US$ 250,000 per person or roughly half the SIGAR rate. According to a U.S. Foreign Relations Committee report (2011) the State Department and USAID are spending around US$ 1.25 billion on capacity-building efforts. 71 from the World Bank and ARTF, to provide the civil service with skilled national and international expertise to carry out technical feasibility studies. But the unit provided mainly short-term inputs and did little to build longer term capacity. In 2002–04 the government, along with ARTF, approved the Afghan Expatriate Program and Lateral Entry Program. (The programs were merged in 2005.) The former program sought to hire a small number of expatriate Afghans as senior advisers, while the latter aimed to address the widespread shortage of competent and experienced civil servants in upper- and middle- management positions in key ministries and agencies, offering “lateral” entry to suitably qualified Afghans. Both programs were succeeded by the MCP in 2007, which aimed primarily to develop a cadre of senior tashkeel civil servants in line ministries, who would undertake reform in departments such as finance, procurement, human resources, and policy and planning. These programs have all helped introduce individual capacity into ministries, but none has been large enough to make a difference at wider government or even ministry levels. The Afghan Expatriate Program recruited only 95 Afghans into the government, and the Lateral Entry Program was scarcely better at 138. The MCP has appointed only 142 Afghans, across 28 ministries and agencies (World Bank 2011a). This scattergun approach to injecting capacity has inevitably diffused its impact. A more targeted and strategic approach to developing the capacity of key ministries is required.44 Where MCP positions have clustered, as in the Ministry of Finance, the impact of MCP appointees has been much greater. But ministerial reform will only occur when there is strong reform-minded leadership in a ministry with enough high-quality MCP-type tashkeel appointees at the top to drive the process throughout the ministry. A handful of MCP positions distributed here and there will not succeed.45 Achieving the right scale of impact will require not only increased, but also long-term, donor support. Second civil service Donor support to ministries is so fragmented that even in a single ministry many human resource departments do not know how many EFS work in or closely connected to that ministry. According to the U.S. Government Accountability Office (GAO 2011b), 587 advisers and mentors were working to build capacity in the Ministries of Interior and Defense alone. A survey by the Ministry of Finance in early 2010 had nearly 7,000 Afghan EFS working in nonsecurity ministries and agencies.46 In this subsection we look at the costs of those in the second civil service who work either directly for government (as short-term contracted civil servants or in largely civil service-type roles in line ministries) or for project implementation units implementing on-budget programs, such as the National Solidarity Program (NSP). We do not consider the much larger universe of individuals who work for contractors, NGOs, and PRTs pursuing mainly off-budget projects. The World Bank collected information from the human resource departments of the eight ministries and one agency that accounted for the largest share of development budget spending in 2011 (table 4.3). In these entities, EFS form 4 percent of the total workforce (excluding teachers), but cost around 44 In his 2009 Atlantic Council article, Ashraf Ghani advocated recruiting up to 3,000 senior civil servants (Ghani 2009). 45 See the conclusions of the Implementation, Completion and Results Report on Afghanistan’s Civil Service Capacity Building Project (World Bank 2011b). 46 This report, “Donor Funded Remuneration of Individuals Working for Government” by the Ministry of Finance, March 2010, collected data mainly from donors rather than ministries’ human resource departments. The results are incomplete as several major donors failed to provide full responses. 72 40 percent of the total on-budget payroll of these ministries (figure 4.1).47 The survey found just under 5,000 EFS working there, all but 200 of them Afghans, estimated to cost US$ 125 million a year. Table 4.3 EFS in key ministries and agencies EFS Ministry/agency Civil service National International Ministry of Agriculture, Irrigation and Livestock 9,156 393 36 Ministry of Education 41,464 1,353 16 Ministry of Energy and Water 2,806 129 24 Ministry of Finance 7,036 237 60 Ministry of Mines 2,709 66 9 Ministry of Public Health 17,750 374 13 Ministry of Public Works 3,378 67 11 Ministry of Rural Rehabilitation and Development 2,110 1,879 21 Independent Directorate of Local Governance 19,971 253 4 Total 106,380 4,751 194 Source: World Bank staff calculations. Note: Ministry of Education data exclude teachers, of whom there are more than 180,000. Figure 4.1 Share of externally funded staff in eight ministries and one agency, 2011 (percent) Workforce Total payroll costs National EFS International EFS 3.9% 0.2% National EFS 31.9% Civil service 60.2% International EFS 7.9% Civil service 95.9% Source: World Bank staff calculations. Salaries for EFS vary among donors and these variations become even more pronounced when allowances and salary top-ups are considered (table 4.4). The IARCSC has been unable to harmonize salaries paid to the EFS. A presidential decree has mandated that the Ministry of Finance and IARCSC should develop such a policy proposal.48 47 Payroll costs of EFS in ministries are calculated from average national and international EFS salaries of US$ 2,200 and US$ 13,300 a month. International salary scales are based on those used by the Civilian Technical Assistance Program in the Ministry of Finance, which are likely to be on the low side for many expatriates. Additional security and life-support costs are excluded. 48 The Commission working on the proposal has been collecting data on technical assistance from donors and ministries for over a year to make recommendations on developing a “harmonized” salary scale acceptable to all parties. There is little evidence that it will complete its task soon. 73 Table 4.4 Salaries paid by donors for similar positions (US$ ) Position World Bank UNDP USAID DANIDA Manager/coordinator 5,134 2,845 4,191 — Senior specialist/manager 2,997 2,367 3,511 2,500 Specialist/officer 1,749 1,709 23,667 1,560 Analyst/assistant 802 852 1,428 770 Support 409 523 790 500 Source: World Bank data. Note: n.a. = not available EFS earn far higher salaries than civil servants: monthly P&G reform salaries average US$ 284 a month, national EFS salaries (paid through World Bank/ARTF-funded projects) average around US$ 2,200 a month—eight times higher. For most of the ministries in the sample, EFS remuneration accounts for nearly 40 percent of the total payroll costs (figure 4.2), and about 90 percent in the Ministry of Rural Rehabilitation and Development (MRRD), which relies heavily on such staff. International EFS earn much more than national EFS. The international EFS in MRRD, for example, account for 0.52 percent of the total workforce but almost 6 percent of the total payroll. In the Ministry of Finance, the national and international EFS constitute 3 percent and less than 1 percent of the total workforce respectively, but their payroll shares are 16 percent and 24 percent. Figure 4.2 Share of EFS in total payroll costs in surveyed ministries and agency 100 International EFS 90 80 National EFS 70 60 Percent Civil service 50 40 30 20 10 0 Ministry of Ministry of Ministry of Ministry of Ministry of Ministry of Ministry of Ministry of Independent Agriculture, Education Energy and Finance Mines Public Health Public Works Rural Directorate of Irrigation and Water Rehabilitation Local Livestock and Governance Development Source: World Bank staff calculations. EFS perform a range of management and technical functions in ministries. Line ministries recruit national EFS to run projects funded by donors (on and off budget). Relatively few are recruited into 74 tashkeel civil service management positions at senior grades, such as through the MCP. Line ministries recruit some to perform specific technical tasks. In the Budget Department in the Ministry of Finance, for example, EFS account for about 66 percent of all staff. Of these only the Director-General of the Department holds a tashkeel position as an MCP placement. Progress on public financial management reforms would never have been achieved without heavy reliance on these specialist contracted staff. In the Ministry of Health, the Grants and Contracts Unit, which handles all donor-funded activities, has no tashkeel civil servants. In MRRD’s national programs like NSP, implementation is almost entirely dependent on contract staff. Six ARTF/IDA-funded national programs across the largest ministries accounted for more than 17 percent of the development budget in 1389 (2010/11), with NSP comprising the major share at nearly 12 percent. At a US$ 50 million payroll cost these programs were the equivalent of 7 percent of the government’s total nonsecurity wage bill in 1389 (2010/11; table 4.5). The government will have no fiscal space to pick up such costs through its operating budget, and will have to continue depending on donors. Table 4.5 EFS employed by key programs and projects EFS Total cost of EFS Allocation for Payroll cost Allocation (annual 2010/11 (percentage of (percentage of US$ (annual US$ total program 2010/11 Program/project National International million) million) allocation) budget) National Solidarity Programme 781 5 21.59 344.00 6.28 11.92 Afghanistan Rural Enterprise 183 3 5.36 10.36 51.65 0.36 Development Program Horticultural and Livestock Productivity 148 11 1.79 16.45 10.87 0.57 Project Public Financial Management Reform 115 3 3.54 14.36 24.66 0.50 Project Strengthening of Health Activities for 90 4 3.04 8.29 36.63 0.29 the Rural Poor Project Education Quality Improvement 532 3 14.64 109.37 13.39 3.79 Project Total 1,849 29 49.94 502.83 9.93 17.43 Source: World Bank data as of June 2011. Since 2002 donors have used the development budget as a way to deliver what is essentially recurrent cost operational support (although attached to development projects) without channeling funds directly through the government’s operating budget. Similar to O&M, these liabilities potentially expose the government to considerable downside risks if donor funds dry up and they are unable to deliver core services (chapters 3 and 5). The MRRD, for example—one of the main development spending ministries, with 19 percent of the share of the 1389 (2010/11) development budget but less than 1 percent of the operating budget—has relatively few regular civil servants. Most of its work is carried out through project implementation units. Implementation is largely through contracted NGOs and other partners, which operate outside the formal civil service and receive payment through the development budget. Conversely, the Ministry of Education employs 222,000 civil servants (including teachers) at a cost of about US$ 341 million, with payroll spending of 62 percent of the ministry’s budget and nearly all its operating budget. It has the largest operating-budget share of any ministry (in nonsecurity), but has a development budget allocation only about half MRRD’s. 75 Despite the frequent criticism of how technical assistance staff are used, and the many recommendations to improve matters,49 neither donors, policymakers, nor individual Afghans have seen much incentive to rationalize their use.50 Part of the reason is that the approach suited everyone. Donors got the best staff for their projects (by skirting the regular civil service) and exercised greater control over these staff by financing these positions either off budget completely or through the development budget. (It was in no single donor’s interest to moderate the salaries it paid, as long as other donors were competing for a limited supply of skilled labor.) Senior Afghans in the ministries benefited from the unregulated environment of salary-supplementation programs, which allowed them to play donors off each other. And the policymakers often had a vested interest in the opportunities for patronage that the system generated. The only body that consistently lost out was the rest of the civil service, which had to depend on P&G to see any salary improvement. The incentive structures that have created the current system are set to change, reflecting a declining aid environment, tightening job market, and greater donor willingness to channel money through government systems (in line with commitments at the London and Kabul Conferences in 2010). Impact of transition The transition offers a compelling opportunity for the government to review its dependency on technical assistance in its development strategy and to consider the extent to which it can afford to absorb current costs in the future. As discussed earlier, over the last 10 years capacity development has been limited, most assistance has been delivered off budget, and Afghanistan continues to rely very heavily on outside staff to run the government and deliver services. Even assistance delivered on budget has often been ring-fenced by project operations that are managed by project implementation units outside the formal civil service and funded through the development budget. Over the next three years a concerted effort needs to be made to bring this capacity within the civil service and government framework. Only in this way can the government regain full control over its own civil service and development programs. Without a mechanism to bring these staff into the service, it risks national EFS withdrawing from such positions, given the uncertainty of transition. Any sharp drop in aid will have serious consequences for both key national programs and the employment prospects of the several thousand skilled Afghans they currently employ. Donors will need to continue paying the costs of these programs for some time, but the government needs to start planning the transition of these programs to their relevant line ministries. Some programs may be unsustainable in the long run without continued donor funding, but cost and efficiency gains could come from bringing these programs under tighter government budgetary control, including lower wage costs and rationalized salary scales. Donors need to support government efforts to harmonize EFS salaries. The current labor market may eventually adjust itself as demand for skilled labor tails off, but donors have a responsibility to support government efforts to retain and attract civil servants. Realistically, this can only be done if donors are 49 At the Afghanistan Development Forum in 2007 various recommendations were agreed upon, such as development of a code of conduct for donors and government to follow in hiring EFS, the need to link capacity building and technical assistance to public administration reform, and the need for all technical assistance to include a training element. Most were never followed up. 50 President Karzai’s directive in January 2010 to set up a Commission to look at the issue of EFS salaries was one of the first indications that the government was taking the issue seriously. 76 prepared to ensure some self-regulation in paying national EFS, and adopt an agreed-upon, harmonized pay scale for such staff. The government itself needs to rationalize civil service pay systems so that it does not undermine its own reform efforts. We estimate the current cost of national EFS in the eight ministries and one agency we reviewed at around US$ 125 million annually (or around 32 percent of their payroll costs). This is an unaffordable cost for the government to absorb, but a fraction of the costs that donors spend on EFS. If technical assistance efforts are to have any impact on long-term capacity, in the future they must be better anchored in the government’s wider capacity-building and public administration reform. A more strategic approach is required to develop capacity and reform public administration, including mechanisms to attract highly skilled staff into the civil service. This approach may still require differentiated or enhanced salaries, but these systems should work within existing government structures rather than parallel to them (box 4.2). The transition will create an impetus for donors and the government to focus on building, rather than substituting for, the capacity of the core civil service in order to create a more fiscally and institutionally sustainable model that delivers basic government services (chapter 5). Box 4.2 Capacity Building for Results Facility Afghanistan’s government and the World Bank have recently signed an agreement to implement the Capacity Building for Results facility. This US$ 350 million five-year facility will be financed through the ARTF, and will provide support to line ministries to facilitate their own reform plans, including the hiring of skilled Afghans from the second civil service directly into the civil service at senior grades. The project draws on two key lessons learned over the past 10 years. The first is that reforms within government must be led by the relevant reforming institution, rather than imposed from outside. This project will facilitate the design of a comprehensive reform plan by individual line ministries seeking to draw on the facility, with the IARCSC and the Ministry of Finance. These reform plans will encourage the development of achievable but challenging service delivery and process reform targets, against which a ministry must demonstrate progress to qualify for funding. The project will ask ministries how they will coordinate ongoing projects within the ministry to achieve these targets, and support their planning for a progressive reduction in donor support to ministry activities. This demand-driven approach will be complemented by support to the ministries. The second key lesson is that capacity building should be grounded on existing ministry structures and should not create parallel institutions. Depending on the ministry’s track record of reform, willingness to participate, and criticality to economic development and service delivery, the facility will offer a package of support to the ministry, including support to key staffing positions. With multiple donor and international agencies competing for qualified staff, the government needs a mechanism to attract high-grade staff back into the core civil service. The facility will finance salaries that are higher than P&G salaries for selected strategic and management positions. The intervention has four main advantages. It will enable the government to offer competitive salaries in a heated labor market, increasing accountability of staff and reducing reliance on EFS. It will help set up a cadre of senior professional staff to drive continued reform of key departments in ministries and agencies. It will work toward institutionalizing civil service management capacity, marking a shift from ad hoc capacity injection to more sustainable management positions. Finally, by emphasizing comprehensive reform throughout ministries, it will stress capacity to enhance service delivery subnationally through better planning, coordination, and use of resources. Source: World Bank. 77 The fiscal implications of shifting at least some of the second civil service into the regular civil service are quite modest relative to current ad hoc capacity-support arrangements. It is estimated that the additional costs to government of paying for a senior civil service group of around 2,000 civil servants (on top of existing P&G salaries) might be around US$ 62 million a year—less than 10 percent of the current total nonsecurity civil service wage bill. (These additional costs are already included in the estimated increase in the overall civilian salaries bill from 5 percent to 9 percent of GDP by 2020/21.) For a poor country like Afghanistan this represents a large fiscal burden on top of an already very large fiscal gap; for the international community, however, it represents a much more efficient way of delivering capacity support than current efforts channeled through many different projects. In the future, with less money flowing into the country, the incentive to build local capacity should become much stronger. As is increasingly clear, the transition will require the government and donors to reappraise how assistance is delivered to citizens through the development and external budgets. Without the security envelope provided by the International Security Assistance Force in the provinces as PRTs draw down, the ability of donors to deploy staff to provincial capitals will become increasingly constrained and potentially much more expensive. Donors will want to use less costly Afghan staff and to deliver services through the government, rather than through parallel systems such as PRTs. While in the short term this desire may affect service delivery, in the medium term it may—if service-delivery programs are prioritized better—mean more efficient and better use of scarce national capacity. EFS will continue to play their part but will be fewer in number and used more selectively. Key national programs, such as NSP, may still be delivered through NGOs, but there may be greater incentive to migrate separate project implementation units into the ministry. Conclusions and recommendations Civil service reforms in Afghanistan have had limited impact on developing the capacity of the core civil service, largely due to insufficient incentives to deal with the second civil service. Although the civil service has evolved since 2002–03 when reforms were initiated—it is larger, more structured, and better staffed—efforts to modernize it through P&G reform have been insufficient to meet the challenge of the far higher salaries in the second civil service. The result is that the government and donors have transferred very few skills into the core civil service. Given the high costs of the second civil service, present modes of service delivery are unsustainable unless donors are willing to continue subsidizing it after 2014. For example, the development budget has substantial salary liabilities for the future operating budget. These costs will be difficult to absorb unless donors continue to fund some of them. The transition will cut the numbers of EFS, which is likely to lead to a short-term fall in service delivery. The government will thus need to prioritize expenditure—it may be unable to support all programs at current levels of delivery. It will also need to adopt a more selective approach to civil service reform, emphasizing key service-delivery and spending ministries, and to focus efforts on selecting academic, managerial, and technical staff on enhanced scales within these ministries. This approach involves important policy choices for the civil service, to be made systematically as part of wider ministry reform, rather than the current ad hoc approach to capacity injection. 78 Adopting a more comprehensive and coordinated approach to capacity development and technical assistance through support to a single ARTF-financed institutional reform program—which focuses on combining salary support, capacity development, and service delivery within a single common public sector reform framework—should secure large efficiency gains for both donors and government. The overriding lesson of the last 10 years is that too many actors and projects chasing too many short-term stabilization—rather than development—goals leads to poor service delivery and an institutional environment that supports quick fixes over longer term capacity development. The government needs to work to retain the confidence of skilled Afghans to stay in the country. But if security deteriorates markedly, no strategy is likely to work as expatriates and skilled Afghans will want to leave. Any sustained capacity flight, as in the 1990s, will severely undermine government efforts to maintain even a modicum of core services. With much less money in the system and less ability to deliver technical assistance through parallel systems after 2014, there should be greater incentive to channel resources through the government. The transition should allow a more disciplined approach to reform, in which the government—in exchange for continued donor support to key public sector salaries—agrees to more demand-driven and results-oriented institutional reform. Eventually, the government should move all core civil service salary support operations from the development to the operating budget. The government needs to use the opportunity of the transition over the next three years to accelerate the transfer of capacity into the core civil service, and to develop delivery, policy, and management capacity in government institutions, aligned with its more straitened resources. These steps require a series of interlinked interventions from the government: confirming policy choices on delivery that take account of even more constrained resources and possibly heightened insecurity; providing reasonable medium-term financing to front-line ministries to improve ministry efficiency and performance; and developing an incentive and governance environment that makes the civil service attractive as long- term employment for Afghans who have acquired high skills in donor programs. In the next chapter we look at the challenges of delivering and maintaining basic services and infrastructure in the face of tightening government and donor budgets. 79 References ADB, UNDP, and World Bank (Asian Development Bank, United Nations Development Programme, and World Bank). 2002. Afghanistan: A Preliminary Needs Assessment for Recovery and Reconstruction. Manila. Congressional Research Service. 2011. “The Cost of Iraq, Afghanistan, and Other Global Wars and Terror Operations since 9/11.” Paper 7-5700. Washington, DC. GAO (U.S. Government Accountability Office). 2011a. “Afghanistan Governance: Performance-Data Gaps Hinder Overall Assessment of U.S. Efforts to Build Financial Management Capacity.” Paper 11- 907. Washington, DC. ———. 2011b. “Afghanistan’s Donor Dependence.” Paper 11-948R. Washington, DC. Ghani, A. 2009. A Ten Year Framework for Afghanistan: Executing the Obama Plan…And Beyond. A report for The Atlantic Council, Washington, DC. Michailof, S. 2007. “Review of Technical Assistance and Capacity Building in Afghanistan.” Paper prepared for the World Bank’s Afghanistan Development Forum, April 26. Ministry of Finance. 2010. “Donor-Funded Remuneration of Individuals Working for Government.” Unpublished Report, Kabul. ———. 2011. Development Cooperation Report, 2002–2010. Aid Effectiveness Unit, Kabul. OECD. 2009. Bridging State Capacity Gaps in Situations of Fragility, Lessons Learned from Afghanistan, Haiti, South Sudan, and Timor-Leste, Partnership for Democratic Governance Experts’ Series. Paris. ———. 2011. Report on International Engagement in Fragile States, Democratic Republic of Timor-Leste. Paris. Rabrenovic, Aleksandar. 2010. Civil Service Professionalization in the Western Balkans. Expert Report for the SIGMA-OECD Project on the State of Civil Service in the Western Balkan States. OECD/EU Brussels Special Inspector General for Afghanistan Reconstruction. 2010. Actions Needed to Mitigate Inconsistencies in and Lack of Safeguards over US Salary Support to Afghan Government Employees and Technical Advisers. Audit-11-5. Washington DC. ———. 2011. The US Civilian Uplift in Afghanistan Has Cost nearly US$ 2 billion and State Should Continue to Strengthen its Management and Oversight of the Funds Transferred to Other Agencies. Audit-11-17. Washington DC. Surhke, A., and Hakimi, A. 2011. “Sustainability and Transition in Afghanistan. A Political Economy Analysis.” Unpublished background paper for this report. World Bank, Kabul UK House of Commons Foreign Affairs Select Committee. 2011. The UK’s Foreign Policy Approach to Afghanistan and Pakistan, Volume 1. Report. London. U.S. Senate Foreign Relations Committee. 2011. Evaluating US Foreign Assistance to Afghanistan. Report. Washington, DC. UNDP. 2007. Practice Note—Capacity Development during Periods of Transition. New York. ———. 2003. Mid-Term Review of Capacity Building Fund, Union of Serbia and Montenegro. New York. ———. 2007. Evaluation Report, Capacity Building Fund, Georgia. New York. 80 USAID. 2008. Audit of USAID/Afghanistan’s Capacity Development Program. Audit Report 5-306-08-012- P. Washington, DC. World Bank. 2004. Sub-national Administration in Afghanistan, Volume 1. World Bank/Afghanistan Research and Evaluation Unit, Washington, DC. ———. 2006. Engaging with Fragile States: An IEG Review of World Bank Support to Low-Income Countries under Stress. Washington, DC. ———. 2008. Afghanistan—Building an Effective State: Priorities for Public Administration Reform. Report 42166-AF. Washington, DC. ———. 2008. Intergovernmental Fiscal Relations and Sub-national Expenditures in Afghanistan, Washington DC. ———. 2011a. Draft Aide Memoire, Management Capacity Program, 9–20 July. Washington DC ———. 2011b. Implementation, Completion and Results Report on Afghanistan’s Civil Service Capacity Building Project. Report ICR00001549. Washington, DC. ———. 2011c. World Development Report 2011: Conflict, Security and Development. Washington, DC. World Bank and UNDP. 2011. International Support for Capacity Development in Post-Conflict States, Reflections from Two Case Studies in West Africa. Washington, DC and New York. 81 Chapter 5 Delivering Services and Maintaining Infrastructure Summary Investments over the past decade, largely financed by aid, have increased access to services and augmented national infrastructure in Afghanistan, but human and social indicators remain among the worst in the world. The country will continue playing catch up in this regard for several decades. Afghanistan’s public finances will be unable to cover the costs of operating and maintaining the infrastructure created in the past 10 years and of delivering the social services and other services currently funded by donors. The government will thus need to focus public resources on essential infrastructure and regular delivery of basic services. This will require hard choices on cutting or shrinking some programs, not maintaining all assets, and managing expectations accordingly. It will have to account for new investments’ operation and maintenance (O&M) funding needs, as well as those for existing key assets and core services. Costs of delivering core public services and operating and maintaining key infrastructure (in the five sectors of basic education, primary health care, agriculture and rural livelihoods, roads, and electricity) should be fiscally affordable, provided aid declines gradually and is shifted on budget. The size and type of risks to service delivery during transition vary by sector, partly due to differences in financing sources and service delivery modalities. Thus sector-specific strategies will be needed to mitigate risks. To assure effective O&M and service delivery, budgetary and administrative reforms are needed. These include developing systems, capacity, and more systematic budgeting and execution of O&M. Introduction Since 2001 the Afghan population’s access to basic services has greatly improved. School enrollment has increased sharply, and as of November 2011 there were 7.8 million children in school, of whom 38 percent are girls. Primary health care coverage expanded from 8 percent in 2001 to 68 percent in 2008. Rural development programs, such as the National Solidarity Program (NSP), have delivered basic infrastructure, including protected water sources, micro-hydropower plants, and improved irrigation to tens of thousands of villages. The country’s road network has been extensively rehabilitated, and travel times between major centers have fallen by up to 75 percent. Rebuilding the electricity sector has also been dramatic, with operational capacity expanding 2.5 times between 2002 and 2011. Still, Afghanistan lags many other low-income countries in social and physical infrastructure. 82 Development activity in Afghanistan is fragmented, with funding spread across many different and uncoordinated off- and on-budget interventions. The government recently sought to introduce greater coherence and prioritize its national development strategy by setting up 22 national priority programs (NPPs; table 5.1). At the July 2010 Kabul Conference, donors endorsed the NPP concept and committed to align at least 80 percent of their assistance along the NPPs and move at least 50 percent of their financial contributions on budget by the end of 2012. Table 5.1 National priority programs Category National priority program Security Afghanistan Peace & Reintegration Program Governance Financial & Economic Reforms National Transparency & Accountability Program Efficient and Effective Government Local Governance Law & Justice for All Human Rights & Civic Responsibility Human resources development Facilitation of Sustainable Decent Work through Skills Development & Market- Friendly Labor Regulation Education for All Expanding Opportunities for Higher Education Capacity Development to Accelerate Implementation of the National Action Plan for Women in Afghanistan Human Resources for Health Agriculture and rural development National Water & Natural Resource Development Program National Comprehensive Agriculture Production & Market Development Program National Rural Access Program Strengthening Local Institutions Infrastructure development National Regional Resource Corridor Initiative National Extractive Industry Excellence Program National Energy Supply Program National Urban Delivery Program Private sector development Integrated Trade and SME Support Facility E-Afghanistan Source: Ministry of Finance. The transition through, and then beyond, 2014—with expected reductions in aid and international military expenditure—will challenge the progress already made in service delivery. The unprecedented recent extension of formal state structures and services to the provinces and districts has mainly been funded by large injections of foreign financial and technical assistance. This has raised the quality and quantity of services, but with serious cost and sustainability implications. The last decade’s boom of exceedingly high foreign assistance has created an unsustainable burden of physical assets and services that the government will find difficult to maintain as donor resources decline. As the transition unfolds, the government will need to recalibrate what is affordable and desirable in size, scope, and implementation of service delivery. With funding for services contracting, it will need to prioritize service delivery, using the criteria of impact, cost-effectiveness, and sustainability. Service delivery expansion since 2001 has been largely program-based, with large shares of funding allocated to flagship national programs, such as the Basic Package of Health Services, NSP, the National Rural Access Program (NRAP), and the National Area Based Development Program (NABDP). These programs, as well as numerous ancillary projects, have been donor financed; all told, nearly 80 percent of the 2011/12 development budget is donor funded. Some key national programs have been implemented by nongovernmental organizations (NGOs) contracted by the government. 83 Sustaining standards of service delivery beyond 2014 will require more funding for O&M. Investments over the past 10 years have raised expectations of continued service delivery that, if not met, could threaten development gains and undermine popular perceptions of government legitimacy. Ensuring continuity in service delivery will require not just funding for civil service payrolls but also higher allocations for O&M and reforms to build the capacities and systems necessary to implement them. Achieving this amid declining donor funding and reduced external capacity will be challenging, but feasible, provided the government is willing to prioritize and rationalize interventions. Government presence and service delivery in three districts A detailed study commissioned by the World Bank on service delivery in three districts—selected to capture some of the variation across Afghanistan’s 364 districts—shows examples of what has happened in the lowest administrative level of government and identifies key transition issues (Altai Consulting 2011a). The three districts represent a wide variation in attributes such as geographic location, proximity to urban areas and markets, and infrastructure connections.51 The study found that in all three districts (Balkh district in Balkh province, Qargaye district in Laghman province, and Yakawlang district in Bamiyan province), basic government administration is present: notably the district governor’s office (with 10 positions in each district, virtually all filled, representing a significant achievement from several years ago) and the Afghan National Police. The presence of other line departments is much more uneven and tends to be more limited, with most administrative capacity at the provincial and central levels. Such strengthening of basic government structures has been accompanied by improved service delivery, as recognized in public opinion surveys conducted for the study. General public expectations have also risen, quite possibly beyond what can be sustainably delivered given declining resources. Some central features of the three districts include the following:  The district tashkeel (number of staff positions) do not match operational capacities (office space, means of transport, and O&M resources to operate and maintain assets).  Only the district governor’s office has consistent support from higher levels of government administration.  Most district administration staff lack transport; the police are the only department with vehicles, so by default they tend to take over other functions (related to justice, for example).  Electricity supply varies greatly by district but has improved in recent years.  They have no procurement capacity.  Village councils and shuras, as well as representative bodies, have proliferated, duplicating some roles.  Municipalities are the only fiscally autonomous administrative level. They have their own revenue sources, and own land and buildings that they can rent out or sell. For service delivery, the study found a varied picture but with common themes (box 5.1). Sectors where district administration has received focused support from higher levels and where services are provided 51 How representative the three districts are, as well as their per capita expenditure in different categories, is uncertain, and even more important, the district expenditure data are probably more relevant for a rural population. About 23 percent of the Afghan population lives in cities and towns where per capita expenditures would reflect the municipalities as well as more provincial and central government spending. 84 under a specific technical framework have relatively high levels of government activity. The three districts have standardized least in agriculture and rural livelihoods, though NSP provides a standard approach to community development, and NABDP at the district level. The districts also vary in their often large road-building and rehabilitation activities. Box 5.1 Key findings on service delivery in three districts Services related to general administration of the districts (such as the district governor’s office and finance) seem to have received focused attention over the past three years for training and operational support. This has helped to integrate and coordinate the district departments somewhat, and to start standardizing administrative and financial processes. Services developed under a specific technical framework, such as education and health, show a higher degree of standardization and equitable distribution across the three districts, even though staff skills need continuous upgrading. The Afghan National Police has received focused support for training and equipment, particularly in Qargaye and Balkh, where local tensions persist. By contrast, the justice institutions have yet to receive the required support and seem to be operating on a very tight budget. All three districts, however, have full representation of the three justice-sector institutions (Ministry of Justice, Attorney General’s Office, and Supreme Court). Construction or rehabilitation of access and productive infrastructure (roads, water, power) is common, though the districts do not plan or cost them (nor does the provincial level). Thus inc luding such assets in the districts’ balance sheets provides an idea of its infrastructure foundation and economic potential, but does not reflect the districts’ decisions or activities. As the passage of a national road through a district may have a very positive impact (Balkh) or a limited impact (Qargaye), the investments linked to such assets may not be the best indicator of equitable resource distribution. Maintenance budgets for these assets are paltry, and will need to be planned at the regional or provincial level. Agricultural and rural livelihood activities are the least standardized and are difficult to quantify and cost. With its historical heritage and existing productive assets, Balkh has received support for developing its agricultural activities, while Yakawlang and Qargaye have benefited mostly from smaller and shorter term aid programs. Without easy access to a strong regional market, Yakawlang has been unable to develop its agriculture beyond subsistence, while Qargaye seems to have no identity of its own, squeezed between a highway and a powerful neighbor (Nangarhar province). The National Solidarity Programme only partly addresses such distortions. The Ministry of Rural Rehabilitation and Development budget is the highest of all ministries’ budgets in Qargaye. Needs for rural water supply remain great and in large part are unaddressed in all three districts. People in all three districts seem to recognize that services have improved. According to user perception surveys carried out in all three districts for this study, a positive correlation is apparent between public opinion on government services and increased access to essential services, which is encouraging. General expectations have also risen, focusing as much on quality as access. These expectations will have to be managed carefully. Source: Altai Consulting 2011a. Despite major data gaps, the study roughly assessed the cost of government functions and service activities in each district, divided into four broad categories (table 5.2):  Governance and essential services (district governor’s office; justice institutions; administrative staff; education, health, and police workers; and O&M costs of sub-national administration and district, education, and health councils).  Other services and development costs (other line departments, whose presence may vary due to local differences, such as Ministry of Rural Rehabilitation and Development (MRRD), the Ministry of 85 Energy and Water (MEW), and the Ministry of Agriculture, Irrigation, and Livestock (MAIL); and development costs in the national development plan, including education, agriculture, and rural livelihoods).  External budget (other than for roads).  Road construction and maintenance (including a mix of core budget and external budget activities— by far the largest share goes to construction and rehabilitation, and only minuscule amounts to maintenance). Table 5.2 Expenditures and costs, three districts, 2010–11 (US$ million, unless otherwise indicated) Total of three National illustrative Balkh Qargaye Yakawlang districts calculation Governance and 3.01 2.60 2.70 8.30 essential services Other services and 1.37 1.23 1.47 4.06 development Total on budget 4.37 3.82 4.17 12.36 1,402 Per capita (US$ ) 47.44 33.99 50.64 43.11 a External budget 4.30 6.84 3.54 15.37 Per capita (US$ ) 46.59 60.84 43.01 53.61 Total excluding roads 8.67 10.66 7.71 27.04 3,065 Per capita (US$ ) 94.03 94.82 93.66 94.27 Road construction and 6.24 1.27 9.08 16.59 maintenance Per capita (US$ ) 67.71 11.30 110.28 57.84 Total 14.91 11.93 16.78 43.63 4,946 Total per capita (US$ ) 161.75 106.12 203.94 152.11 Source: Altai Consulting 2011a. Note: Illustrative national totals are derived by applying per capita figures for the total of the three districts to the estimated national population figures. The population figures for 2008 are: Balkh, 92,200; Qargaye, 112,400; Yakawlang, 82,300; and national, 32,517,656. (Because population estimates for the districts are available only for 2008, the figures are grossed up using the 2008 national population estimate from the World Development Indicators.) a. Based on data from the United States and rough estimates for other donors’ assistance outside the Afghan national budget. Data for the three districts and Afghanistan’s population were used to derive illustrative estimates of national costs for government presence and service delivery among districts. The resulting figures, shown in the right-most column of table 5.2, need to be interpreted with considerable caution, but suggest:  The rough estimate for on-budget spending (that is, excluding roads and the external budget) grossed up nationally (US$ 1.4 billion)—around 9 percent of 2010 gross domestic product (GDP), or 82 percent of domestic revenue—seems within fiscally viable resources, especially as revenue is projected to grow robustly.  Adding roads and the external budget makes district spending unsustainable nationally—at roughly US$ 5 billion a year, this is equivalent to 33 percent of 2010 GDP and more than three times domestic revenue. Road construction is heavy in two of the three districts (and presumably in many others). It is likely to decline as aid falls in the coming years, raising issues of maintenance and funding. 86 The in-depth data-gathering and analytical work on the three districts confirms the picture suggested by the broader national data in earlier chapters: it will be fiscally impossible in a worsening aid environment to operate and maintain the many assets and services built up in a fragmented manner over the past decade. Stringent prioritization is required. Service delivery in five key sectors Among the NPPs, five sectors are particularly important for securing economic growth and meeting basic needs: education, health, agriculture and rural livelihoods, roads, and electricity (figure 5.1 below gives recent core government budget data). Each sector has experienced tremendous growth over the past 10 years, in funding and outputs. But the sectors differ widely in how they deliver services, modes of financing expenditure, and the funding they command from on- and off-budget sources. This section reviews the current situation and the issues faced in sustaining basic service delivery in these sectors, including funding requirements for O&M. Education Enrollment in primary and secondary schools increased from 1.1 million in 2001 to 7.2 million in 2011.52 Net school enrollment climbed from 43 percent to an estimated 68 percent for boys and 3 percent to 44 percent for girls. In 2011, 170,000 students were enrolled in government-sponsored Islamic schools; 21,000 attended 61 technical and vocational training schools; and 550,000 (including 330,000 women) attended literacy courses. Enrollment in tertiary institutions also increased significantly, from 22,500 in 2002 to 68,000 in 2011. The 2010–11 Afghan government budget allocated US$ 706 million to education: US$ 514 million operating costs and US$ 193 million development spending (figure 5.1). Salaries for the Ministry of Education and the Ministry of Higher Education totaled US$ 454 million (93 percent for teachers’ salaries). Allocations for purchases of goods and services totaled US$ 56 million. The construction and rehabilitation of schools, universities, and other facilities received US$ 104 million. Off-budget education expenditure reached at least US$ 128 million in 1389 (2010/11).53 52 Sources of information for this subsection include World Bank (2011a,b), Ministry of Education (2010 and 2011a,b), Education Management Information System (eMIS), and Ministry of Finance budget statistics. 53 Off-budget expenditure data were collected from Canada, Germany, Japan, the Netherlands, USAID, and provincial reconstruction teams. 87 Figure 5.1 Core education budget by expenditure type, 1390 (2011/12) (US$ million) Curriculum development $6 million Miscellaneous Vocational education $10 million $10 million Literacy programs Teacher education $13 million $33 million Capacity building $16 million Construction $104 million Assets (Op.) $4 million Goods and services (Op.) $56 million Salaries (Op.) $454 million Source: Ministry of Finance 2011a. Education is delivered primarily by civil servants paid from the operating budget (its share is higher than that of any other sector except security). The bulk of the Ministry of Education’s payroll is made up of teachers, most of whom have undergone pay and grading reform. This reform, as well as the introduction of merit-based hiring and promotion, has boosted teacher salaries to an average US$ 1,360 a year. With recurrent spending concentrated in the operating budget, donor contributions are directed primarily to new infrastructure, providing the sector with relatively high levels of overall funding. Education assets have mirrored the growth in student numbers, and the number of schools rose from 6,039 in 2002 to 13,643 in 2011. School construction has crowded out O&M, however, with O&M allocations falling far below requirements and rarely reaching schools. Opaque procedures discourage local officials from applying for the few funds available, and no systems appear to be in place for higher authorities to assess requests. Lack of O&M funding has caused education infrastructure to deteriorate: school officials surveyed in late 2010 from the three districts discussed in the previous section (Balkh, Yakawlang, and Qargaye) reported a lack of school furniture and the need for urgent repairs. Education receives extensive development funding from the Afghanistan Reconstruction Trust Fund (ARTF) and USAID (figure 5.2). ARTF contributed some 50 percent of the sector’s development budget in 1390 (2011/12), including most notably the Education Quality Improvement Project. Afghan government funding accounted for US$ 29 million of the 1390 (2011/12) development budget (89 percent for building schools, universities, and other educational facilities).54 USAID contributed US$ 10 million to the core education budget and spent an additional US$ 83 million off-budget on community-based 54 Other donors with on-budget contributions include Denmark (US$ 26 million), World Bank (US$ 23 million), and USAID (US$ 10 million). 88 education, the American University of Afghanistan, the American School of Kabul, skills training for Afghan youth, and small grants to education-focused NGOs.55 Figure 5.2 Education development budget by donor, 1390 (2011/12) (US$ million) Other bilaterals $3 million Other $4 million USAID World Bank $10 million $23 million ARTF $97 million Denmark $26 million Government of Afghanistan $29 million Source: Ministry of Finance 2011a. Despite rapid gains in student numbers and the high current expenditure, access to education remains low in rural areas, particularly for girls. In 2010 enrolled students were estimated to be only 58 percent of the total school-age population, and almost 15 percent of those enrolled are permanently absent. Variations across provinces in education outcomes are wide, with primary net enrollment rates ranging from 30 percent in some areas to almost 100 percent in others. The variation in girls’ primary enrollment is especially wide, reportedly ranging from less than 10 percent in Helmand and Uruzgan to almost 90 percent in Nimroz province.56 The current school population is also heavily concentrated in grades 1–4, with high dropout rates in higher grades. Widespread concerns exist over education quality, owing to the poor qualifications of some teachers, lack of a standardized curriculum, lack of evaluation standards, and insufficiency of basic school supplies. Statistics collected in 2009–10 by a national teacher registration system indicate that only 27 percent of the 162,000 registered general education teachers are educated at a grade 14 level57, the 55 Other large donors supporting off-budget education programs are Germany (US$ 13 million), Canada (US$ 12 million), and Japan (US$ 11 million). According to Ministry of Education estimates, around US$ 250 million is spent on education annually through the external budget. 56 Girls’ school enrollment in many areas appears to be constrained not so much by cultural or attitudinal factors, but by the long and often unsafe walks that students have to make to reach the nearest school and associated security and other concerns. 57 two years of postsecondary education 89 official minimum requirement for teaching, or higher. Previous studies have also shown leakages in systems supplying school materials, with schools surveyed across three provinces in late 2010 reporting receipt of only a fraction of those requested (Altai Consulting 2011c). Health Public health coverage has improved.58 From 2002 to 2007, the number of functioning primary health care facilities increased from 496 to 1,169, and the proportion of facilities with female staff rose from 25 percent to 83 percent. In 2001, only 8 percent of the population was estimated to have access to basic health care, but in 2007/08, more than 68 percent lived within one hour’s walk of a basic health facility. Infant mortality fell from 14.8 percent of live births in 2002 to 13.4 percent in 2009, and childhood immunization rates climbed from 12 percent of children ages 12–23 months in 2005 to 37 percent in 2008. Coverage remains patchy, however. For example, in the two remote central provinces of Ghor and Daykundi, only 14 percent and 17 percent of households live within one hour’s walk of a basic health facility. Skilled birth attendance in 2008 was only 24 percent (but up from 16 percent in 2005). Even in provincial centers, tertiary facilities are rudimentary at best. Public facilities in rural areas, where they exist, are open irregularly, and medical professionals commonly request extra payments or refer patients to private practices or pharmacies in which they have a financial interest. Public health spending focuses on the Basic Package of Health Services (BPHS). This was established by the Ministry of Public Health (MoPH) in 2003 to deliver high-impact primary health interventions, now covering all rural districts, with services delivered through a hierarchical network of facilities. Complementing BPHS is the Essential Package of Hospital Services, which standardizes hospital services and promotes a referral system that integrates BPHS facilities with the hospital network. Government capacity in service delivery is very limited, however, and NGOs employ 90 percent of rural health workers and half the provincial hospital staff. To productively exploit this external capacity built during the 1990s, when Afghanistan lacked a formal health policy, BPHS was designed to be carried out by NGOs. Apart from the three provinces of Kapisa, Panjshir, and Parwan, BPHS is still delivered by NGOs contracted either directly by MoPH or by the European Union, which currently channels funding off budget. Contracted NGOs are responsible for training service providers and for staffing facilities, as well as ensuring that facilities are adequately supplied and maintained. Of the NGOs contracted by MoPH, 27 percent are international; nearly all the NGOs contracted by the European Union are international. Most delivery capacity therefore remains outside the government: 60 percent of the country’s 20,000 registered health workers are employed by NGOs and most of the rest work at MoPH central and provincial offices. Public spending is relatively low. The 1390 (2011/12) budget allocated US$ 201 million to health—US$ 57 million operating and US$ 144 million development spending (figure 5.3). BPHS accounted for 55 percent of development budget spending in 1390 (2011/12). Funding for that budget comes mainly from USAID and the Center for Disease Control (CDC; 50 percent), the World Bank (16 percent), and the Afghan government (14 percent; figure 5.4). An additional US$ 125 million off-budget funding was reported by major donors, with US$ 41 million from USAID and the CDC and US$ 30 million from the European Union. Total funding in 2010/11 was estimated at US$ 326 million, of which 30 percent went to BPHS. 58 Information in this subsection is from the World Bank and the Ministry of Public Health, as well as the Ministry of Finance (budget statistics). 90 Figure 5.3 Core health budget by expenditure type, 1390 (2011/12) (US$ million) Miscellaneous $12 million Capacity building Salaries (Op.) $9 million $37 million Construction $15 million Disease control $11 million Goods and services (Op.) $19 million Essential Package of Hospital Services Asset acquisition (Op.) $19 million $1 million Basic Package of Health Services $79 million Source: Ministry of Finance 2011a. Figure 5.4 Health development budget by donor, 1390 (2011/12) (US$ million) Other bilaterals $15 million GAVI / Global Fund $13 million Government of USAID / CDC Afghanistan $72 million $20 million World Bank $23 million Source: Ministry of Finance 2011a. 91 High private health expenditure supplements the low public outlays. Public health spending rose steeply from US$ 3.72 per capita in 2001 to US$ 10.94 in 2009 and is now comparable with low-income country averages (WHO 2011).59 But public spending is low relative to the population’s needs, and the private sector serves a disproportionate role in health care provision, with out-of-pocket payments to health care providers accounting for three-quarters of total health spending (Ministry of Public Health 2011). As such payments are often very large, high private spending implies that financial distress often hits affected households. Despite recent improvements, health outcomes remain inadequate. Afghanistan is short of medical professionals. And due to their unwillingness to practice outside major cities, this shortage disproportionately affects rural areas—where most of the population lives. Thus, even with the progress made in the past few years, health service coverage remains thinly spread, and the country still suffers from very poor health outcomes. It is, for example, now one of only four countries in the world with endemic polio. Agriculture and rural livelihoods Rural livelihoods interventions, particularly NSP, have commanded substantial funding in recent years, partly because Afghanistan’s population is overwhelmingly rural—80 percent of the population is estimated to live outside regional and provincial centers. The country’s topography, unreliable rainfall, and history of civil conflict mean that those living in rural areas suffer from tenuous livelihoods and limited access to basic amenities such as clean drinking water, reliable irrigation, and electricity (see below). To improve the quality of village-based infrastructure, increase the economic security of the rural population, and enhance government legitimacy in rural areas, MAIL, MRRD, and MEW have carried out numerous donor-funded interventions. The most prominent rural livelihoods intervention is NSP, which, since 2003, has disbursed block grants averaging over US$ 33,000 to 28,397 communities in all 34 provinces to help create and rehabilitate village infrastructure.60 Community development councils—created by NSP and elected through universal suffrage with secret balloting—manage the grants, and 8 national and 21 international NGOs support implementation. The program is in its third phase, which will create community development councils and disburse grants in the roughly 16,000 communities that have yet to benefit from the program, and will extend “repeater grants” to the 12,000 communities that have fully used their first grant. NSP spent US$ 217 million in 1389 (2010/11) and has US$ 208 million in the 1390 (2011/12) MRRD development budget (figure 5.5). By 2015, program spending is expected to cumulatively total US$ 2.5 billion. Funding for other rural livelihoods interventions is dispersed across numerous small interventions pursuing various objectives. In the 1390 (2011/12) development budget, MAIL listed some 28 development projects with an average budget allocation of US$ 3.2 million. Of these, 10 focus on agricultural productivity, three on irrigation, five on agribusiness, eight on livestock, and two on forestry. The largest MAIL program is the US$ 100 million semi-autonomous Agriculture Development Fund, 59 Total per capita health expenditure in Afghanistan, which includes private spending, rose from US$ 11.12 in 2001 to US$ 50.89 in 2009. Average public and total health expenditure for low income countries in 2009 were US$ 11.66 and US$ 28.53, respectively. Average public and total health expenditure for South Asia (excluding Maldives) in 2009 were US$ 23.57 and US$ 45.88, respectively. 60 Of block grants disbursed by NSP from 2003 through mid-2011, cumulatively, 26 percent have been spent on transport infrastructure, 24 percent on water supply and sanitation, 19 percent on irrigation, 12 percent on power, and 10 percent on schools. 92 sponsored by USAID, which loans funds directly to agribusiness, and through lending by financial and nonfinancial institutions to farmer cooperatives and associations for buying agricultural inputs. The largest non-NSP MRRD project is the US$ 50 million Afghanistan Rural Enterprise Development Program, which seeks to build the managerial and financial capacity of rural microenterprises. No other MRRD project for rural livelihoods exceeds US$ 4 million. Figure 5.5 Core rural livelihoods budget by expenditure type, 1390 (2011/12) (US$ million) Miscellaneous Goods and $1 million services (Op.) $8 million Salaries (Op.) MEW irrrigation $27 million projects $56 million Capacity building Assets (Op.) $12 million $0.5 million MAIL projects $89 million NSP $208 million MRRD projects $31 million Source: Ministry of Finance 2011a. The development and external budgets for rural livelihoods are large (among the five sectors). In addition to the US$ 432 million allocated through the core budget, US$ 474 million in external budget expenditure was reported in 1389 (2010/11) by major donors. On-budget funding comes primarily through the ARTF, the World Bank, Afghan government resources, the United States, and the Asian Development Bank (ADB; figure 5.6). The largest off-budget donor is the United States, which spent US$ 361 million in 1389 (2010/11) on improving agricultural productivity, regenerating agribusiness, rehabilitating watersheds, and improving irrigation infrastructure. The next largest off-budget donor, the European Union, contributed US$ 66 million to a range of programs on food security, agricultural production, and river-basin management. 93 Figure 5.6 Rural livelihoods development budget by donor, 1390 (2011/12) (US$ million) Other $7 million Other bilaterals $44 million ADB $32 million ARTF World Bank $199 million $39 million Government of Afghanistan $39 million USAID / USDA $37 million Source: Ministry of Finance 2011a. NSP, which provides block grants to villages for small-scale projects prioritized by elected Community Development Councils, has been one of the most successful national programs. It has reached some 26,000 villages, and more than 45,000 sub-projects have been completed, in every province and district of Afghanistan. When NSP was designed, the initial block grants were deliberately set at a high per- capita level, reflecting the long neglect of rural development, urgent local reconstruction priorities, higher costs of construction materials, and in particular the need for government to make a significant impact in rural areas to enhance its legitimacy. According to NSP data the average value of block grants is US$ 39 per capita, compared with averages for similar programs in other countries of US$ 1.5–US$ 2.0 in India, US$ 1.5 in Indonesia, US$ 5.5–US$ 26.0 in the Lao People’s Democratic Republic, and US$ 4– US$ 5 in the Philippines.61 61 Given lack of data on village populations, and perhaps some confusion about the definition of a “household” in the Afghan context, it is possible that the NSP figure may be understated. 94 Roads Through outlays of some US$ 4 billion, Afghanistan’s road network has been much improved.62 Travel times between Kabul and Kandahar have fallen from 19 hours to 5 and between Kabul and Herat from 44 hours to 12. Road rehabilitation since 2001 has covered all different types of roads. The most prominent work has been on the national ring road, which connects Kabul, Kandahar, Herat, and Mazar- e Sharif. After this and other projects to rehabilitate primary roads connecting provincial centers were largely completed, investments have focused on rehabilitating secondary and tertiary roads.63 Although some have been carried out off-budget by PRTs and bilateral donors, the National Rural Access Program (NRAP)—jointly run by the Ministry of Public Works (MPW) and MRRD—has served as the primary means for rehabilitating rural roads. International and local private contractors, through MPW, MRRD, or the donor directly, implement road construction, rehabilitation, and resurfacing on secondary and tertiary roads. Works on tertiary roads are generally carried out by local contractors through competitive bidding process and community-based contracting through direct contracting with community development councils. Core budget funding for transport (including the entire MPW budget and the portion of the MRRD development budget for NRAP) is mainly for road construction, either through NRAP or individual projects. It totaled US$ 345 million in 1390 (2011/12; figure 5.7). Of this, around US$ 200 million went to full scale rehabilitation of the road network and US$ 11 million to partial road rehabilitation and resurfacing. NRAP, which includes allocations to both MPW and MRRD, had funding of US$ 66 million. (Rail received US$ 27 million.) ADB is the primary donor to the development budget, followed by ARTF, the Islamic Development Bank, the World Bank, and the United States (figure 5.8). USAID also provides much off-budget funding for road construction, rehabilitation, resurfacing, and maintenance through USAID’s Infrastructure Rehabilitation Program. 62 Information sources for this subsection include USAID (2010 and 2011a,b) and Altai Consulting (2011a). 63 Primary roads include national and regional highways and urban roads; secondary roads all provincial roads; and tertiary roads all rural roads. 95 Figure 5.7 Core transport budget by expenditure type, 1390 (2011/12) (US$ million) Miscellaneous $2 million Goods and services, Railways including O&M (Op.) $27 million $28 million Salaries (Op.) Assets (Op.) $10 million O&M $0.1 million $2 million National Rural Access Program $66 million Rehabilitation / resurfacing $11 million Construction $200 million Source: Ministry of Finance 2011a. Figure 5.8 Transport development budget by donor, 1390 (2011/12) (US$ million) Other $0.1 million Other bilaterals $48 million ADB $101 million USAID $20 million World Bank $22 million Islamic Development Bank $25 million Government of Afghanistan ARTF $53 million $37 million Source: Ministry of Finance 2011a. 96 Maintenance underfunding has noticeably degraded the quality of roads built since 2001. The 1390 (2011/12) budget establishes a separate US$ 27 million fund for road maintenance, though there is no systematized mechanism to designate priorities or to verify the quality of works contracted out. As the next section explores, this amount is far too little to meet the road network’s maintenance needs. Current maintenance activities are generally implemented off budget on limited stretches of road. One of the most successful maintenance programs in recent years has been the Infrastructure Rehabilitation Program, which has financed maintenance for 3,500 kilometers of roads built by USAID and ADB. It employs a performance-based management model that assigns sections of road to local firms, which assume responsibility for routine and emergency maintenance to ensure a pre-established level of service. In 2009, more than US$ 200 million was collected in transport-related revenue in the country, mainly fuel and vehicle import duties, licensing, and transit fees. But this revenue was not earmarked for transport and was thus reallocated through the general revenue fund. To create a sustainable basis for future maintenance and expansion of the road network, USAID has proposed setting up a national road authority through a public–private partnership to centralize revenue collection and contract maintenance works using a performance-based model (similar to that used by the Infrastructure Rehabilitation Program). Electricity Protracted conflict by 2001 had devastated the country’s ability to generate, transmit, and distribute electricity.64 In 1978 Afghanistan had around 396 megawatts (MW) of generating capacity available to the grid, but by 2002 this had fallen to 243 MW. Investments since then—particularly in the past few years—increased this capacity to 948 MW by June 2011. From August 2006 through August 2009, electricity production more than doubled from 888,040 MWh in 2006 to 1,804,150 MWh in 2009. MEW estimates that, as of 2010–11, 30 percent of the population have access to electricity, with large economic centers such as Kabul and Mazar-e-Sharif having access to 24-hour power for the first time in decades (Da Afghanistan Breshna Sherkat 2011). Increases in electricity access in Kabul and much of northern Afghanistan have been driven by the North East Power System, which built transmission systems to import 550 MW of power from Central Asia. 65 Progress has also been made in rehabilitating distribution systems for the major cities. Efforts to rehabilitate and expand domestic generation capacity have been less successful, however, and 71 percent of operating capacity in June 2011 came from imported power (48 percent) and diesel plants (23 percent). Rehabilitation of Kajaki dam in Kandahar, for example, faced numerous setbacks due to lack of security, though it is now delivering 18 MW to Helmand and Kandahar provinces. Installation of the 105 MW Tarakhil plant near Kabul was also delayed and ran heavily over budget, and because of high diesel prices its generating costs are six times that of imported electricity, so it runs at only 6 percent of capacity. Afghanistan has huge hydroelectric potential,66 but insecurity, drought, logistical challenges, and cross- border riparian issues have hindered efforts to rehabilitate or build hydroelectric plants. Hydroelectric 64 Information sources for this subsection include the Ministry of Energy and Water (2007), Da Afghanistan Breshna Sherkat (2011), Inter-Ministerial Commission for Energy (2010 and 2011), and Altai Consulting (2011a). 65 Similar investments in western Afghanistan have improved transmission and distribution systems, linking urban centers to sources in the Islamic Republic of Iran and Turkmenistan. 66 A potential generation capacity of 23,000 MW has been cited. Southern Afghanistan has suitable conditions for efficiently generating solar power, and western Afghanistan for generating wind power. 97 sources account for 20 percent of operating capacity, thermal for 5 percent, and renewable sources for 4 percent. As the cost of diesel generation (and even imported electricity) is higher, failure to fully exploit hydroelectric resources results in a heavy financial burden.67 To reduce technical and nontechnical losses, improve efficiency, and limit political interference, the national power utility (DABS) was corporatized in 2008. It has since begun modernizing its power management technologies and adopting monitoring procedures.68 DABS revenue grew from US$ 141 million in 2009/10 to US$ 172 million in 2010/11, slightly exceeding consumption increases over the same period. USAID is planning to directly contract the construction of US$ 1.4 billion in generation and transmission infrastructure to DABS, and to build long-term human resource capacity in key areas. Allocations for O&M—in the development budget and in the goods and services component of the operating budget—are limited (figure 5.9). While revenue can theoretically cover such recurrent costs, collection rates are too low to do this. The managerial and technical capacity of MEW and DABS is also a concern, as both entities continue to experience difficulties in attracting and retaining qualified staff. Although DABS is experienced in operating the existing 110 kilovolt (kV) system, it does not have the same experience with the 220kV/500kV transmission systems being built or with large-scale generation assets. So O&M of the system depends heavily on foreign staff provided by the U.S. Army Corps of Engineers and private contractors. Figure 5.9 Electricity core budget by expenditure type, 1390 (2011/12) (US$ million) Miscellaneous Goods and $1 million services (Op.) $2 million Salaries (Op.) $5 million Assets (Op.) Capacity building $0.1 million $6 million O&M Generation $2 million $23 million Distribution $39 million Transmission $51 million Source: Ministry of Finance 2011a. 67 The cost of imported electricity is US$ 0.06/kWh (Zorpette 2011), the cost of diesel generation is US$ 0.20–US$ 0.46/kWh (Special Inspector General for Afghanistan Reconstruction 2011), and the cost of hydroelectric generation is estimated at US$ 0.03/kWh (Trofimov 2011). Electricity is sold domestically at around US$ 0.08 per kWh. 98 The complexity of off-budget activities and the recent commercialization of DABS make it difficult to obtain an accurate picture of the scale and composition of outlays in the electricity sector. A database maintained by the Inter-Ministerial Committee on Energy lists US$ 1.6 billion in current and US$ 6.3 billion in planned donor projects through 2019. The United States (including the U.S. Army Corps of Engineers and U.S. forces) is the largest donor to current projects (US$ 660 million), followed by Germany (US$ 345 million), India (US$ 204 million), ADB (US$ 136 million), and ARTF (US$ 114 million). The planned project database is dominated by the United States (US$ 5.7 billion) and ADB (US$ 566 million). Figure 5.10 Electricity development budget by donor, 1390 (2011/12) (US$ million) Islamic Development Bank $2 million World Bank $5 million ARTF $15 million ADB $47 million USAID $20 million Government of Afghanistan $32 million Source: Ministry of Finance 2011a. Even with recent growth in consumption, the country has a large unmet demand for electricity. By some estimates, with per capita electricity consumption as low as 21 kilowatt-hours per year, Afghanistan consumes less electricity per person than any other country worldwide. With high economic growth and increasing urbanization, demand can be expected to grow rapidly in coming years. This will inevitably create pressures that, at a time of declining external funding and technical capacity, will present onerous technical and management challenges to DABS and MEW. Sustaining service delivery The five sectors show widely differing delivery modalities, but have made few efforts to develop sustainable funding. Agencies external to the government have made the most investments to expand service delivery (except in education). The delivery of BPHS and NSP, for instance, is almost entirely 99 reliant on NGOs, and new investment and maintenance in transport and electricity have generally been undertaken by private contractors and funded off budget. In addition, resources are concentrated in the development budget (except for education; figure 5.11), making it hard to institutionalize sustainable capacity for service delivery. The following section outlines the transition’s implications for delivering the five key services. To provide an indication of the funding that will be required to sustain the delivery of services at current levels, the section presents estimates of the O&M expenditure necessary to sustain current infrastructure in a useable state.69 As detailed in the tables in the previous section, current O&M expenditure is very low. Significant increases in O&M funding in many of the sectors will be required to avoid deterioration in infrastructure and the delivery of key services. Figure 5.11 Core government budget, 1390 (2011/12) allocations and 1389 (2010/11) expenditure (US million) 800 700 600 Development budget 500 $ million O&M 400 Salaries 300 200 100 0 2011/12 2010/11* 2011/12 2010/11* 2011/12 2010/11* 2011/12 2010/11* 2011/12 2010/11* Education Health Rural livelihoods Transportation Electricity Source: 1389 expenditures are from Ministry of Finance (2011b), with salary and O&M expenditure estimated using shares of operating expenditure sourced from the 1389 national budget. 1390 allocations are from Ministry of Finance (2011a). Note: Due to the negligible expenditure on the asset allocation code of the operating budget, the code is excluded from the totals. Although service delivery across the sectors is at risk from overall declining aid, loss of capacity, and potential security threats, it also faces sector-specific difficulties due to diversity in implementation modalities and levels and sources of funding. 69 It is important to note that such O&M expenditure is in addition to the recurrent costs of employing service providers (such as teachers and medical staff) and ministry staff. 100 Education Education may be relatively protected from transition-linked pressures. NGOs and private contractors play a small role, and the operating budget, which covers teacher salaries, is funded mainly from domestic revenue and the ARTF, offering greater predictability in funding flows. The heavy O&M requirements, however, threaten to disrupt service over the medium term. The overall education budget is large, but the proportion of funding for O&M—such as books, furniture, and building maintenance—is very small. Combined with the lack of an effective system to identify and prioritize needs and disburse assigned funds accordingly, this has resulted in a failure to adequately attend to O&M of educational facilities. If this approach persists, standards of education services will deteriorate. The failure to adequately provide for O&M is especially acute in view of the scale of construction activities, which account for US$ 104 million in the 1390 (2011/12) development budget and relate to an additional 1,920 schools, 15,000 computers, and 75,000 educational laptops designated for acquisition by 2014. Based on current plans for asset construction and acquisition, it is estimated that US$ 234 million in O&M expenditure will be required annually by 2014 to sustain Ministry of Education assets alone (table 5.3). This figure contrasts sharply with the 1390 (2011/12) budget allocation to the Ministry for non payroll O&M of just US$ 30 million, of which only a small share appears to reach provinces and districts. The extent to that local communities, PRTs, NGOs, and off-budget donor interventions cover O&M needs is unknown. Table 5.3 O&M needs for education infrastructure (US$ million) Year Desks Chairs Computers Vehicles Supplies Facilities Books Laptops Total 2011 8.8 0.6 0.1 1.5 56.8 102.2 0.0 0.0 170.0 2014 9.7 0.7 0.4 1.7 98.9 119.8 0.1 2.5 233.8 Source: Altai Consulting 2011b. O&M activity is constrained not only by funding, but also by requisitioning issues, such as the absence of budget earmarking for local, district, and provincial use and a lack of due diligence among higher officials in processing requests. Thus any increase in the O&M budget needs to be accompanied by earmarking of funds and a communication campaign for school officials on how to access these funds. The demographic bulge—34 percent of the population is under 10 years of age—is expected to generate continuing rapid growth in student numbers over the coming years, requiring school construction as well as teacher training and hiring. The education budget is forecast to grow sharply: the wage bill alone is projected to reach US$ 607 million by 2014. And if current demographic trends continue, by 2020 an additional US$ 495 million will be required to accommodate another 4 million primary and 3 million secondary students. Such high and growing costs may well give rise to fiscal pressures to rationalize costs. Health In contrast to education, health services are delivered predominantly by NGOs and funded exclusively from the development and external budgets. Moreover, just two donors fund much of health service delivery: the United States and the European Union. A sharp drop in financing from donors (particularly the United States), or even worse their withdrawal, would seriously affect the rural population’s access to health care. 101 The annual cost of supplying and maintaining Afghanistan’s public health system is expected to be US$ 78 million by 2014 (table 5.4). But if programs are not expanded and funding remains at current levels, the O&M estimates would be lower, at around US$ 45 million. The O&M costs of BPHS facilities are factored into the value of contracts between MoPH and NGOs, so the costs of operating and maintaining assets ranging from basic health centers to mobile clinics are included in the development budget. Assuming that the current NGO-based implementation modality for BPHS continues, the only items that need to be added to the operating budget are the O&M costs of regional and special hospitals and vehicles. Table 5.4 O&M needs for health infrastructure (US$ million) Basic Comprehensive Regional/ health health District Provincial Mobile national Special Year centers centers hospitals hospitals Subcenters clinics hospitals hospitals Vehicles Total 2011 4.7 10.8 11.4 7.5 1.6 2.9 0.9 1.7 3.1 44.5 2014 8.0 18.2 19.3 12.7 2.7 4.9 1.6 2.9 7.2 77.5 Source: Altai Consulting 2011b. Any disruption to health service delivery would be acutely felt. Because of low funding, BPHS services are thinly spread, but they deliver essential interventions such as child immunization, tuberculosis and malaria control, maternal and newborn care, and medications. Any disruption or reduction could deny life-saving measures to vulnerable pockets of the population, worsening the country’s already dire health indicators. Agriculture and rural livelihoods Activities in rural livelihoods are also vulnerable to transition pressures. Sector interventions are implemented primarily by NGOs (NSP) and for-profit contractors (other projects), assisted by either private consulting firms or directly by Development Partner organizations (such as GIZ), and are overwhelmingly funded through the development and external budgets. Rural livelihoods interventions other than NSP tend to be relatively small, with funding sources for each project concentrated in single donors. NSP (a large program with numerous donors) nonetheless may be vulnerable to funding reductions by virtue of its disproportionately large share of budgetary allocations and the pressures this generates, particularly when overall resources decline. As in other sectors, underfunding of O&M threatens service delivery standards. NSP has, for example, funded construction of 10,400 water and sanitation facilities and 4,500 micro-generators since 2003. Although Community Development Councils managing these assets are supposed to designate plans and funding to meet O&M requirements, it is not clear how well this works in practice. Despite limited evidence, there may also be issues of inadequate O&M for irrigation facilities constructed or rehabilitated by MAIL, MRRD, and MEW, not directly falling under the purview of village communities. The overall O&M needs of rural livelihoods assets in 2011, excluding village-level infrastructure built under NSP and other community-based programs,70 is estimated at just US$ 8 million a year. The O&M cost will rise to just shy of US$ 11 million a year by 2014 once the currently planned assets are created (table 5.5). 70 According to NSP, community development councils are required to prepare and implement O&M plans for infrastructure subprojects completed in their community. As a result, the O&M liabilities of these assets, which do not fall under the central government’s purview, are not included here. 102 Table 5.5 O&M needs for rural livelihoods infrastructure (USUS$ million) Irrigation Year Wells Dams Canals Total 2011 3.0 0.1 3.5 7.6 2014 11.3 0.1 2.5 13.9 Source: Altai Consulting 2011b. O&M needs for canals is likely under-estimated. Roads The transition threatens to reduce the available funding for roads, but also brings a sharper focus on sustainability. Due to systematic underfunding for maintenance over the past 10 years, a good part of the country’s 39,000 kilometer (km) road network needs full rehabilitation, but this will be extremely expensive—much more than pursuing a regular pattern of maintenance would have been. Routine maintenance of a kilometer of primary road, for example, costs an estimated US$ 3,000 every year, with periodic maintenance every four years of US$ 10,000—but full rehabilitation costs US$ 420,000. Current estimates suggest that the sector will need to invest around US$ 3 billion in rehabilitation to put the country’s road network in a maintainable state. The cost of maintaining the road and bridge network is estimated at US$ 290 million in 2011. By 2014, the network is scheduled to grow by 7,000 km to 46,000 km, pushing maintenance costs to US$ 394 million a year (table 5.6). These figures include routine, periodic, and emergency maintenance activities but not the enormous rehabilitation requirements noted above. Table 5.6 O&M needs for road infrastructure (USUS$ million) Year Primary Secondary Tertiary Bridges Total 2011 169.0 29.9 52.7 38.3 289.9 2014 180.0 155.7 58.7 47.4 394.4 Source: Altai Consulting 2011b. Developing sustainable funding and pushing through road maintenance will be critical given the heavy maintenance requirements and unpredictability of donor commitments beyond 2014 (and the need to strictly prioritize declining external funding). The government has recently committed to establishing a national road authority, which could centralize transport-related revenue collection and contract maintenance works using a performance-based model. In the meantime, sufficient O&M resources must be allocated to prevent further deterioration (the additional allocations to MPW in the 1390 budget fall well below likely requirements). Electricity The last 10 years of investments have greatly expanded people’s access to reliable electricity, but, as with roads, have also brought O&M liabilities that present sustainability challenges. Some sources of generation capacity, particularly diesel, are expensive, and illegal connections and leakages lose tens of millions of dollars in revenue annually. The sector therefore receives very large subsidies from donors and budgetary revenue for operational costs and capital investment. The corporatization of DABS is an attempt to make the sector more efficient and more independent from external sources of funding. The sustainability of investments in the energy sector will largely depend on DABS’s success in increasing revenue collection, shifting generation to low-cost sources, and developing the managerial and technical capability necessary to undertake maintenance activities. This last challenge is particularly important because external entities such as the U.S. Army Corps of 103 Engineers will likely scale back involvement during the transition, and because MEW and DABS previously had difficulty attracting and retaining qualified technical staff. The O&M cost of electricity infrastructure is put at US$ 194 million in 2011 (table 5.7), with an additional required outlay of US$ 26 million to establish camps to maintain transmission lines. Estimating O&M requirements in 2014 is difficult due to uncertainty regarding future demand, input costs, operationalization of additional facilities, and the resulting relative shares of different generation sources. Assuming completion of the planned expansion of transmission and distribution facilities but no construction of new generation assets, O&M requirements in 2014 could be as little as US$ 297 million with an additional outlay of US$ 52 million to set up more transmission-maintenance camps. But if the additional hydroelectric plants are built as planned, annual O&M could rise to as much as US$ 448 million by 2014. For comparison, DABS reported revenue of US$ 174 million in 2010/11 and is achieving cost-recovery ratios of 75 percent. Table 5.7 O&M needs for electricity (US$ million) Year Diesel fuel Diesel plants Imports Hydro plants Substations Transmission Distribution Total 2011 48.1 5.5 87.2 19.1 4.2 24.8 5.6 194.4 2014 48.1 5.5 109.8 19.1 12.4 84.7 17.1 296.5 Source: Altai Consulting 2011b. Investments to boost energy independence need to take account of costs of electricity provision and avoid high-cost solutions. Ensuring the financial sustainability of Afghanistan’s energy sector will require reducing revenue losses and using less expensive generation sources. Given that electricity can be bought much more cheaply from neighboring countries, efforts to build domestic generating capacity through large-scale diesel generators are fiscally unsustainable. Domestic hydroelectric generation has low operating costs but high capital costs and long gestation periods for investment, so hydroelectric would be integral for any longer term solution to electricity needs. Summary of estimated O&M costs for all five sectors The sector estimates above suggest total nonwage O&M requirements in 2014 of a little more than US$ 1 billion (table 5.8). This is fiscally not affordable unless more efforts are undertaken to recover costs for electricity and transportation services. Table 5.8 Summary of transition challenges for five key sectors (US$ million) Education Health Rural livelihoods Transport Electricity Primary implementation modality Civil service NGOs NGOs Contractors SOEs Core budget allocation, 2011/12 706 201 432 345 182 Salaries and wages 454 37 27 10 21 O&M (non payroll) 56 19 8 28 39 Development 193 144 397 306 121 Donor contribution to 2011/12 development 163 124 358 253 89 budget Estimated O&M needs in 2014 234 77 11 394 297 Shortfall (2011/12 O&M allocation versus 204 67 3 317 258 2014 needs) Main sector-specific threats to service NGO or donor NGO withdrawal Lack of Revenue Lack of O&M delivery withdrawal Lack of O&M O&M losses Source: Tables 5.3–5.7 and World Bank staff assessments. Note: Totals may not sum because of rounding and exclusion of asset acquisition costs portion of operating budget. 104 Conclusions and recommendations Given declining aid and intensifying fiscal pressures during transition, the government will need to consolidate priorities and funds around delivery of core basic services and infrastructure to protect key development gains and support economic growth. Public finances will be unable to meet the costs of operating and maintaining all the infrastructure assets created in an often fragmented manner over the past 10 years or of delivering the social services and other services currently funded by donors. The sheer scale of development needs and the large amounts of aid from many disparate sources has encouraged a scattering of interventions responding to a multitude of priorities. The development of NPPs represents an attempt to impose greater coherence on development interventions, although given the expected deep falls in funding and the rising pressures on the budget from additional O&M liabilities, the scope of activities proposed by the NPPs may be unrealistic. Sustaining minimum standards of basic service delivery will require additional resources, programmatic restructuring, contingency planning, and, in some cases, major institutional reform. Given the pressures of transition, a greater focus on the challenges of sustaining core services is warranted. This will require difficult decisions on cutting or shrinking some programs, not maintaining all assets, and managing expectations accordingly. The government will need to focus public resources on ensuring regular delivery of basic services and maintaining essential infrastructure. The cost of delivering a core set of public services and operating and maintaining key infrastructure should be fiscally affordable, provided aid declines gradually and shifts on budget. The size and type of risks to service delivery during the transition vary by sector, partly because of their financing sources and service delivery modalities. Thus sector-specific strategies will be needed to manage and mitigate sector-specific risks, beyond more general risks of declining overall funding and changing security threats. In programmatic planning and budgetary allocations over the past 10 years, policymakers have for the most part preferred to invest in building new assets rather than maintaining old ones, and in some areas, initial gains in service delivery have been undermined as assets have degraded and fallen out of use. As stakeholders continue to construct new facilities, the liabilities need to be accounted for and funded (as well as those for existing infrastructure). Without adequate provision for O&M, the impact of further development investments will be limited and short-lived. Addressing O&M needs requires not just an increase in funding, but also comprehensive reforms to budgetary and administrative practices. No government ministry currently has the mechanism, institutional capacity, or asset registers to appropriately budget for O&M on a unit cost basis or to ensure that allocations are deployed as intended. The reforms should establish regularly updated asset inventories with descriptive information on conditions and service requirements for each; O&M costing schedules with unit costs, depletion rates, and labor requirements; a formal system allowing field asset managers to relay information on unscheduled maintenance requirements; a rule-based method to prioritize O&M activities; and regular auditing of O&M spending to minimize misappropriation. 105 References Altai Consulting. 2011a. Strengthening District Level Accountability and Service Delivery in Afghanistan: Mapping of Fund Flows in three Pilot Districts. Kabul. World Bank ———.2011b. Assessing O&M of Key Assets in Afghan Provinces. Kabul: World Bank. ———.2011c. Public Expenditure Tracking Survey—Education. Kabul: World Bank Beaston-Blaakman, A., and M. Cros. 2011. A Cost-Efficiency Analysis of Two Alternative Models for Implementing the Basic Package of Health Services in Afghanistan. Kabul: Delegation of the European Commission to Afghanistan. Da Afghanistan Breshna Sherkat. 2011. Energy Consumption and Available Energy Resources in Afghanistan. Kabul: Da Afghanistan Breshna Sherkat. Inter-Ministerial Commission for Energy. 2010. Energy Sector Status Report: July–September 2010. Kabul: Ministry of Economy. ———. 2011. Energy Sector Status Report: April–June 2011. Kabul: Ministry of Economy. Ministry of Education. 2010. Education Summary Report 1388. Kabul: Ministry of Education General Directorate of Planning and Evaluation. ———. 2011a. Education for All—Draft for Consultation. Kabul. ———. 2011b. Education Interim Plan 2011–13 (Version 5). Kabul. Ministry of Energy and Water. 2007. Power Sector Strategy for the Afghan National Development Strategy. Kabul. Ministry of Finance. 2011a. 1390 National Budget. Kabul: Government of the Islamic Republic of Afghanistan. ———. 2011b. Annual Fiscal Report 1389. Kabul: Government of the Islamic Republic of Afghanistan, Fiscal Policy Unit. Ministry of Public Health. 2011. Special Inspector General for Afghanistan Reconstruction. 2010a. Afghanistan Energy Supply Has Increased but An Updated Master Plan Is Needed and Delays and Sustainability Concerns Remain. Kabul. ———. 2010b. Afghanistan Energy Supply Has Increased but An Updated Master Plan Is Needed and Delays and Sustainability Concerns Remain. Kabul. ———. 2011. Trofimov, Y. 2010. “U.S. Rebuilds Power Plant, Taliban Reap a Windfall.” Wall Street Journal, July 13. USAID (U.S. Agency for International Development). 2010. “Officials to Establish Road Authority.” Retrieved September 26, 2011, from http://afghanistan.usaid.gov/en/USAID/Article/1304/Officials_to_Establish_Road_Authority. ———. 2011a. “Road Operation and Maintenance Capacity Building Project.” Retrieved September 26, 2011, from http://afghanistan.usaid.gov/en/USAID/Activity/214/Road_Operation_and_Maintenance_Capac ity_Building_Project. ———. 2011b. The Road Authority—Road Fund Model. 106 World Bank. 2011a. Afghanistan Country Overview 2011. Washington, DC. Retrieved September 20, 2011, from http://go.worldbank.org/1238X2WDR0. ———. 2011b. World Development Indicators. Washington, DC. WHO (World Health Organization). 2011. NHA Indicators. Retrieved September 23, 2011, from http://apps.who.int/nha/database/PreDataExplorer.aspx?d=1. Zorpette, G. 2011. “Struggling for Power in Afghanistan.” New York Times, July 6. 107 Chapter 6 Conclusions and Recommendations Afghanistan faces serious challenges in coming years. Transition will be far more complex than is implied by the current widespread focus in public pronouncements and the media on the drawdown of international military forces and assumption of leadership and full responsibility in this area by Afghan forces in 2014. A broader and longer term perspective on transition is essential. While 2014 is expected to be a significant milestone, it is likely only one among several over the past decade and during the next two. This report has analyzed some dimensions of transition that will be important determinants of its success and development outcomes from a medium- to longer term perspective. These include:  Economic impacts: growth prospects, structural change, employment, poverty, and future economic drivers (chapter 2).  Fiscal challenges: expenditure liabilities, large fiscal gaps and how to close them, revenue prospects, and absorption of more aid on-budget (chapter 3).  Government capacity and how to boost, consolidate, and sustain it during transition: rationalizing the “second civil service” and transforming it into sustainable core government capacity, and continuing public administration reforms (chapter 4).  Implications for service delivery: focused on education, health, agriculture and rural livelihoods, transport, energy, and government functionality and spending at district level (chapter 5). In all of these areas, the distortions created by Afghanistan’s high aid dependency and reliance on the international community, as well as the impact of the coming drawdown of international troops (with associated reductions in international military expenditures) and the anticipated decline in aid, have been assessed. Lessons from history—Afghanistan’s and other countries’—have been drawn upon for the analysis (chapter 1). This chapter briefly highlights some of the main findings in earlier chapters and then discusses key recommendations. Main findings Political uncertainty and insecurity could undermine Afghanistan’s transition and development prospects. International experience (summarized in World Development Report 2011: Conflict, Security, and Development) demonstrates that violence and especially protracted internal insurgency are extremely damaging to development, and that political stability and consolidation are key ingredients of transitions to peaceful development. This underlines ongoing efforts to reach a peaceful solution to the Taliban insurgency, and the urgent need for better political consolidation, particularly in the run-up to the next election cycle. (The presidential election is to be held in 2014, and parliamentary elections in 2015.) But if there is worsening insecurity and increasing uncertainty about longer term stability, Afghanistan’s development prospects will be harmed. The extremely high level of annual aid—roughly US$ 15.7 billion in 2010, about the same as Afghanistan’s gross domestic product (GDP)—cannot be sustained. Aid has funded essential public services under government auspices (implemented directly or through government-contracted NGOs), 108 including education, health, infrastructure investments, and government administration. And there have been substantial improvements in the lives of Afghans over the last 10 years. But these inflows, most outside the Afghan budget, are so high that waste, corruption, aid dependency, and parallel systems to circumvent the government’s limited absorptive capacity have impeded aid delivery and building a more effective Afghan state. The public spending that these aid flows have financed—both on and off budget—will be fiscally unsustainable once donor funds decline. Less aid with more effective aid delivery could, in the end, lead to more positive outcomes. The key is how to manage declining aid, mitigate the adverse impacts, and put aid and spending on a more sustainable long-term path. International experience and Afghanistan’s history after the Soviet military withdrawal in 1989 demonstrate that violent fluctuations in aid, especially abrupt cutoffs, are damaging and destabilizing. Large financial inflows outside the Afghan budget and fragmented aid under weak governance have been major sources of rents, patronage, and political power. This situation has sometimes inadvertently heightened grievances and conflicts as the relative strength of elite groups in Afghan society shifted. As aid declines, reliance on the opium economy and other illicit activities for political economy purposes could increase. And such pressures on government budgetary revenues and expenditures are likely to grow. The need to ensure that increasingly constrained public funds, including domestic revenues and on-budget aid, are used properly reinforces the importance of maintaining the progress by the Ministry of Finance in establishing public financial management systems and supporting a robust Afghan budget process. The impact of declining aid on economic growth will be less than expected. Why? Because most international spending “on” Afghanistan is not spent “in” Afghanistan, and much of what is spent in Afghanistan leaves the economy through imports, expatriated profits, and outward remittances. Even so, projections suggest that, under reasonably favorable baseline projections, real GDP growth may fall from 9 percent a year in 2000–10 to 5–6 percent over 2011–18. Given Afghanistan’s annual population growth of 2.8 percent, there would be limited increases in average per capita incomes, continuing high underemployment, and little progress in reducing poverty. Only growth at the upper range of plausible scenarios would enable Afghanistan to significantly reduce poverty and increase average per capita incomes. For example, with real GDP growth of 6 percent a year, average per capita income—currently among the lowest in the world at US$ 528—would take 22 years to double, about a generation. Economic growth would be even lower under less favorable scenarios. Growth projections are based on assumptions (scenarios) related mainly to security, sources of growth, aid levels, and changes in the investment climate. If the assumptions in the less favorable scenarios come to pass—for example, if agriculture performance is poor, major mining investments (Aynak for copper and Hajigak for iron ore) do not materialize, or aid declines precipitously over the period—then growth could drop to 3–4 percent a year. Deteriorating security and governance would lead to even lower or possibly negative economic growth. While the Kabul Bank crisis has had a negligible macroeconomic impact, and banking sector problems are not expected to adversely affect future growth, the underdeveloped financial sector and low financial intermediation leave little scope to help Afghan businesses adjust to slowing growth. The decline could be partly mitigated by reducing aid in a gradual, planned manner and by increasing the amount of aid actually spent in Afghanistan—for example by channeling more aid through the Afghan budget (on-budget aid has a much higher impact on the Afghan economy than off-budget assistance does). Underemployment will rise because the activities affected by declining financial inflows (services, construction) are relatively labor-intensive. Roughly 6–10 percent of the working population has benefited from aid-financed jobs, though most are short term. So declining aid can be expected to heighten underemployment (with fewer casual labor opportunities and lower pay for skilled 109 employees), even if unemployment is not greatly affected. The adverse impact of lower economic growth on employment is likely to be much larger than the direct employment impact of declining aid, making this a concern during transition. Lower aid and public spending will affect some groups more than others. Aid has not been evenly spread across the country. Because of donors’ choices, and the predominant role of stabilization and military spending, conflict-affected provinces have had much higher per capita aid than more peaceful (and often poorer) provinces. As a result, the slowdown in aid will be felt more acutely in conflict-affected provinces and urban centers. If aid declines gradually so that it can be partly offset by growth of the mining and civilian public sectors, the impact could be softened and spread over time, allowing labor markets more time to adjust. The direct poverty impact of declining international spending might be limited if aid is more equally distributed across provinces, and if assistance shifts toward development programs rather than short- run stabilization activities. Aid disproportionately devoted to conflict-affected provinces has only modestly impacted poverty. Moreover, households in the conflict-affected provinces were less poor on average to begin with, so this concentration of aid inadvertently increased inequality among provinces and between groups. National programs delivered through the government, such as the Basic Package of Health Services and the National Solidarity Program, have been much more balanced across provinces, benefitting Afghans more equitably. The most severe adverse impact of transition will be on the fiscal situation—Afghanistan will face a projected financing gap of 25 percent of GDP by 2021/22, even higher in some of the intervening years. The fiscal gap will be enormous despite hoped-for robust growth in domestic revenue (projected to rise from 10 percent of GDP to more than 17 percent a decade from now). This gap arises primarily due to high wage and nonwage security costs (projected at around 17.5 percent of GDP in 2021/22), along with funding the government and sustaining service delivery on the civilian side. The civilian wage bill is projected to increase to 9 percent of GDP, the non-security operation and maintenance (O&M) expenditure to 4 percent, other operating spending to 2.5 percent, and the core development budget to 10 percent. To close this financing gap, Afghanistan must rely on continuing international funding to pay for most security costs, as long as the size of the security sector (Afghanistan National Army and Afghanistan National Police) remains at or close to recently agreed targets (352,000 personnel with an estimated sustaining cost of US$ 5 billion a year). And civilian aid will need to be spent more selectively for development and O&M. Various foreign aid and domestic resources are possible for different spending categories. But a reasonable approach could be for domestic revenues to pay for the full civilian operating budget and part of the security costs (at the current level of 3 percent of GDP). Donors would finance the remaining bulk of security costs plus a more highly prioritized core development budget. Comparable low-income countries receive around 9 percent of gross national income in non-security development assistance, whereas Afghanistan would require close to three times that. If these levels of foreign assistance for security and civilian expenditures are not available, then the government will need to make extremely difficult and likely destabilizing tradeoffs—either grossly underfunding or significantly shrinking Afghan security forces or crowding out essential civilian spending, or both. Public resources will need to be focused on ensuring regular delivery of basic services and essential infrastructure. Investments in national programs over the past decade, almost entirely funded by donors, have increased access to basic health and education, rural access, power, and irrigation. Even so, Afghanistan’s development indicators remain poor. The country’s public finances will not be able to absorb the costs of operating and maintaining the infrastructure assets created in an often fragmented manner over the past 10 years and delivering the social and other services financed through donor- 110 funded programs. The Afghan government will only be able to afford a modest package of basic social and infrastructure services. Even this will require difficult decisions: cutting programs, not maintaining all assets, and managing expectations accordingly. Better outcomes would result if sustainable institutions for maintaining key infrastructure are developed and operated transparently. Strict priorities among the many demands for O&M and services will also be needed to provide fiscal space for critical public investments required to sustain growth. Increasing on-budget aid and managing O&M through government systems would improve aid effectiveness, but the government will need to overcome serious absorptive capacity constraints to be able to receive and effectively use additional donor money on budget. After core development spending more than doubled in absolute terms between 2005/06 and 2007/08, disbursements have stalled at just under US$ 1 billion annually over the past four years. This is mainly a result of unrealistic budget formulation, large budget carryovers from previous years, budget rigidity resulting from earmarked donor funding, and limited government capacity to implement projects on time. The execution of the operational budget has generally been very high (96 percent in 2010/11), but almost all on wage spending. In the future, managing nonwage O&M—expected to increase from US$ 335 million to US$ 4.8 billion by 2015/16 (of which US$ 1.3 billion would be for civilian O&M)—will be a major challenge. Enhancing the core civil service, with an emphasis on strengthening budget execution and service delivery, will be crucial for government functioning and providing essential services. The heavy reliance on a “second civil service” of externally funded Afghan staff needs to be transformed into a reliance on core government capacity—not to directly deliver all services but to oversee service delivery. Transition provides an opportunity to rationalize the terms and conditions of externally funded staff, bring them into a coherent framework of government capacity, and integrate them with the regular civil service over the longer term—all putting government functioning and service delivery on an affordable, sustainable path. Even then, the government will not have enough resources to absorb the costs of these externally funded staff. And some foreign-provided technical assistance will need to remain. Thus donor financing for externally funded staff and technical assistance will need to continue, but in a way that better aligns with the government’s planned rationalization and strengthens its core capacity. Ensuring the delivery of services to the Afghan people requires delegating more responsibilities to the provincial level. Only a tiny fraction of the O&M budget gets outside the line ministries in Kabul. An important priority moving forward will be enhancing the capacity of provincial offices to participate in budget formulation and key spending ministries to execute their budgets subnationally. Without this, it could be difficult for the government to absorb a greater proportion of aid on budget and deliver results to its people. 111 Appendix 1 MAMS for Afghanistan: Model Structure and Simulations Methodology: Overview of the MAMS framework MAMS is an economy-wide simulation model created to analyze development strategies and their impact on economic growth.71 The model integrates a relatively standard dynamic recursive computable general equilibrium model with optional modules that link government policies and other relevant indicators to poverty and other outcomes for the Millennium Development Goals. By contrast to partial models, MAMS takes into account intersectoral flows in the economy. To what extent and how the model is disaggregated depends on the questions raised, the specific country context, and data availability. Below we describe the overall model as applied to the case of Afghanistan, referring to the Afghani database that we constructed. We give more detailed explanations of the model when describing the BASE scenario in the next section. Among the Millennium Development Goal indicators, we consider poverty but not others indicators because of data limitations. Figure A1.1 gives an overall illustration of MAMS for Afghanistan. The actual model takes the disaggregation further, however, such as different commodity and activity sectors and different production factors, and divides the donor and government sectors into security and nonsecurity. Figure A1.1 Aggregated payment flows for MAMS, Afghanistan Wages and rents Factor Private savings markets Households Transfers Taxes Private net lending Factor costs Indir. taxes Government consumption Government Private consumption Transfers Intermedi. input costs Donor consumption Donors Private investment Activities financing Transfers FDI Rest of the Commodity markets World RoW net Sales lending Imports Exports Donor investment Government investment Private investment 71 This section draws on Bourguignon and others (2008). For detailed technical documentation of MAMS see Lofgren and others (2012). 112 Source: Authors’ illustration. The core computable general equilibrium model includes a government institution (on budget), the services of which are split into two functions: security and nonsecurity, with each of these requiring current (consumption) and capital (investment) spending. Like other production activities, these government sectors use production factors and intermediate inputs to produce an activity-specific output (in the case of the government this means different types of services). In addition, there is a donor institution that also provides public services but with a budget external to the government, fully managed by donors. The donor institution also provides services disaggregated into two functions: security and nonsecurity, both requiring consumption and investment spending. The combination of government and donor sectors is referred to as the public sector. The private sector is divided into opium, agriculture, mining, other industry, and private services. The factors of production in the model include labor, public capital stocks by government activity, and a private capital stock. The government finances its activities from domestic taxes, other domestic revenue, domestic borrowing (expected in the future), and foreign aid (borrowing and grants). The donor sector is fully financed by aid grants. Growth in the stock of public (government and donor) nonsecurity capital contributes to overall growth by adding to the productivity of other production activities (both private and public); growth in the security capital stock generated by public investments is only partly linked to productivity.72 An assumed share of the mining capital adds to public capital, reflecting the benefit of mining-related infrastructure for the rest of the economy. The model includes a single household (an aggregate private domestic institution). The receipts of the household consist of factor incomes, transfers from the government, and transfers from the rest of the world (including worker remittances). These receipts are allocated to direct taxes (reflecting government policy), savings (with a savings rate that responds to higher per capita incomes net of taxes), transfers to the government, and consumption (using demand functions derived from utility maximization). In its interactions with the rest of the world, Afghanistan spends foreign exchange on imports and on interest, and how much varies among the household, government, donors, and firms in the different sectors. This implies that the domestic impact of activities will differ. The country’s receipts stem from exports; foreign transfers to the government, the donor sector, and households; government borrowing from abroad; and foreign direct investment. Domestic commodity supplies are imperfectly transformable between two destinations—domestic sales and exports; the ratios between sales to these two destinations respond to changes in the ratio between domestic sales and export prices. Similarly, commodity demanders view imports and domestic output as imperfect substitutes and respond to relevant relative price changes. A MAMS country database is a synthesis of information from a variety of sources, structured to meet the requirements of the model. The model parameters are then defined using this data. The main components of the database are a social accounting matrix (SAM) and other data that reflect the functioning of the economy. More specifically, the information is primarily related to stock and share data (for labor and other production factors) and elasticities (related to substitutability in production, consumption, and trade as well as to responses in employment and poverty indicators). For the simulations, it is also necessary to provide assumptions about the evolution of policies and other factors that are exogenous to the model. 72 Some security capital is strictly military related and assumed necessary to keep the same level of security. But some security capital involves investments in roads that will add to the economy’s productivity. 113 For the Afghanistan MAMS database, several sources were used, including various publications by the Government of Afghanistan, International Monetary Fund, and the World Bank, covering macro and fiscal data, sectoral studies, enterprise surveys, and household surveys. Data for other countries considered similar in terms of relevant structural features were used when the required information was unavailable for Afghanistan. This should not be seen as a drawback of the analysis. Parameters that have been established in the literature have been used that fall within the range of those for similar countries. Nevertheless, given data weaknesses, the simplifications that are inherent in any model and the economic context in a fragile state such as Afghanistan, the results should be taken as approximate indicators of the effects of alternative policy options, to be considered in conjunction with other inputs to economic policy making in Afghanistan. Simulations Using MAMS, we carried out a set of simulations to assess the effect of alternative policies and exogenous shocks on the macroeconomic and sectoral evolution of Afghanistan’s economy from 2010/11 to 2025/26, but results for the transition period 2010/11–2018/19 will also be reported for some of the indicators. The BASE scenario was designed to reflect a likely development of the Afghan economy while, in each non-BASE scenario, selected assumptions were altered. Each scenario is briefly described in table A1.1, and further under each section below. The assumptions and results of the BASE scenario will be extensively described while the sections on alternative scenarios will focus on how the assumptions and the simulated results differ from BASE. Note that MAMS and its database include the opium sector in its GDP numbers and that the figures in the reference year for 2010/11 are adjusted for historical averages.73 Table A1.1 Descriptions of scenarios Scenario Description BASE Expected developments, including mining expansion (Aynak and Hajigak) and gradual decline in aid Alternative agricultural scenario AGR– Same as BASE but constrained growth in the agricultural sector Alternative aid scenarios AIDALLOC Same as BASE but increasing share of aid on budget AID– Same as BASE but rapid decline in both security and nonsecurity aid Alternative mining scenarios MIN+ Same as BASE but additional mining expansion (Aynak, Hajigak, and 11 other mines) MIN– Same as BASE but no mining expansion Alternative security and governance scenario GOV– Same as BASE but drastic decline in the security situation (investment effects), aid, and no mining development Source: World Bank Base scenario, BASE The BASE scenario (BASE) is designed to represent a plausible development for Afghanistan, incorporating data and projections from multiple sources. It serves as benchmark for comparison with 73 In addition, the MAMS database has been adjusted to ensure consistency (in the sense that the receipts and outlays of different sectors and institutions are equal), a prerequisite for accurate statistics that rarely (if ever) has been imposed on Afghani economic statistics. 114 the alternative scenarios, which address the consequences of major structural changes in the economy, including the scaling down of aid, the allocation of the aid that remains, mining developments, as well as the level and composition of fiscal expenditure. The basic assumption for the BASE is that, unless otherwise stated, the main variables of the economy develop in a manner that is similar to recent trends with little structural change. In selected areas, such as changes in aid and the mining sector, we incorporate additional information. In the BASE, real GDP growth is exogenously determined for 2011/12–2013/14 in order to reflect IMF medium-term projection but is otherwise endogenous for the remaining years.74 The difference in GDP growth rates between the baseline and the following scenarios results from a change in other assumptions, which are specified for each scenario. For the BASE, real GDP growth at factor cost grows at an average of 5.9 percent over 2011/12–2018/19 and 4.9 percent over 2011/12–2025/26.75 In this scenario, among other things, we assume a positive trend in tax revenue, a gradual phasing out of both security and nonsecurity aid, and development of the two mines Aynak and Hajigak. The mining assumptions used in the BASE scenario (and hence all other scenarios apart from the scenario with additional mining expansion, MIN+) are based on mining data from the World Bank/IFC (2011a) study. It includes the exploitation of the copper mine Aynak and the ore mine Hajigak. The MIN+ scenario will include another eleven potential mines. Production for Aynak is assumed to start in 2015/16 with 100,000 ton and then reach 200,000 ton in 2016/17. Investments will total 2,500 million dollars over the period 2012/13–2018/19. For Hajigak production is assumed to start with 5 Mtons in 2016/17 and reach 10 Mtons in 2017/18, with total investments of 2,500 million USD over the period 2013/14–2019/20. The price of copper is estimated to US$ 8875 per ton, and the cost of production to US$ 4450 per ton, while the price of iron ore is estimated to US$ 160 per ton and production cost to US$ 75 per ton. Profits are then calculated and amortization taken into account, resulting in government revenue (royalties, corporate income tax on profits, and bonuses) and net profits. Assumptions of imported goods for investments and intermediate goods, the share of foreign workers and remittance of net profits affect the domestic impact of mining activities. For the (on-budget) government revenue side, domestic nonmining tax rates are increased gradually and the tax rates on the mining operating surplus are set to replicate exogenous projections for the tax intake from the mining sector. Domestic transfers to the government (representing miscellaneous domestic nontax receipts) are a fixed share of absorption. These assumptions will almost double domestic revenue as a share of GDP for the projected period. Foreign grants to the government budget are exogenous in foreign currency and decrease gradually during the whole period. Foreign borrowing is exogenous in foreign currency, set to let the government foreign debt grow at the same rate as GDP. On the spending side, government consumption and investments in both the security and nonsecurity areas grow with constant budget shares, while the shares of security and nonsecurity spending themselves changes over time. Figure A1.2 presents the shares in 2010/11—for the government 60 percent is used for nonsecurity (out of which 34 percent is investments) and 31 percent for security (almost exclusively consumption)—and figure A1.3 present the security and nonsecurity spending developments over time. 74 Productivity growth is in part endogenous (depending on public capital stocks) and in part exogenous. In 2010/11, the otherwise exogenous part is adjusted to generate the target growth rate. Opium and mining have sector-specific treatments. For opium, productivity growth rates are scaled down on the assumption that the sector is discriminated against in terms of laws and policies. For mining, total factor productivity (TFP) growth is not explicit to ensure that production replicates exogenous data. 75 A large share of the base year data is based on IMF data but the difference in results can be referred to: the MAMS includes the opium sector in GDP, which grow slower than the average and corresponds to about 17 percent of GDP; external aid is larger initially and aid in general is assumed to decline faster in MAMS; and a lower TFP growth is assumed in MAMS. 115 Total government expenditures are determined by the limits of the fiscal space (total receipt net of domestic transfers from the government to the private sector). Figure A1.2 Expenditure shares in the government and donor budgets, 2010/11 (percent) Government Donors 100 Other transfers 9.3% Civil 90 investments 25.7% 80 Civil investments Civil 70 39.4% consumption 12.3% 60 Security Percent investments 50 18.6% Civil consumption 40 20.5% Security 30 investments 0.2% Security consumption 20 Security 43.4% consumption 10 30.5% 0 Source: Authors’ calculations. 116 Figure A1.3 Assumed on-budget aid and external aid developments in the BASE scenario (gradual decline), 2009/10–2025/26 (2009/10 USUS$ billion) 9 8 Donor, security 7 6 2009/10 $ billion 5 Donor, nonsecurity 4 3 2 Government, nonsecurity 1 Government, security 0 2009/10 2011/12 2013/14 2015/16 2017/18 2019/20 2021/22 2023/24 2025/26 Source: Authors’ calculations. The donor sector (covering external budget public receipts and spending) has a separate budget (see figure A1.2). All revenues come from external grants which decline gradually as illustrated in figure A1.3. In foreign currency the 2025/26 level is 15 percent of the level in 2010/11. The donor spending allocation is assumed to gradually increase the share of nonsecurity spending and by 2025/26 all external security aid is phased out. Of the total donor budget, in 2010/11 about 42 percent were security related and 58 percent nonsecurity (see figure A1.2). Of nonsecurity spending about 68 percent is investment, while about 30 percent of the security budget is investments. In the private or household sphere of the economy, incomes are determined by factor returns, transfers from government (a fixed share of absorption) and transfer from the rest of the world (exogenous in foreign currency with growth at the same rate as GDP). These incomes are allocated in roughly fixed shares to direct taxes, transfers, consumption, and savings. The nongovernment domestic savings are driven by a savings rate that depends on income net of direct taxes and adjusts (relatively marginally) in response to changes in per capita net income as well as an exogenous component, which is marginally increased during the period on the assumption that the investment climate will improve somewhat.76 Private investment is driven by available financing from private savings and FDI (net of government domestic borrowing, which are assumed to be zero in the case of Afghanistan). FDI, split into mining and nonmining, is exogenous in foreign currency. 76 Improvements in the general investment climate will be discussed below but is not least expected given the development of resource corridors from mining investments. 117 The markets for private capital factors are cleared by its rent. Over time, supply is endogenously determined for nonmining private capital (and labor as described below), but exogenously for other factors (mining capital and land, split into land for opium and land for other agriculture). The marginal product of public (government and donor) capital is initially set at 14.2 percent,77 and will after that depends on depreciation and the growth in sectors which profit from the accumulation of new capital— that is, private nonmining and, to less extent, government sectors. This corresponds to projected internal rates of return of 5–6 percent.78 To consider possible capacity constraint when the government takes over parts of the external budget, or just as a general analysis of the importance of efficiently manage public investments, there will be an alternative scenario assuming a 50 percent lower marginal product of public capital. We now turn to the results of the BASE scenario. As shown in table A1.2, the resulting average annual growth rate for GDP at factor cost for the BASE scenario is 4.9 percent per year for 2011/12–2025/26, and for the transition period, 2011/12–2018/19, 5.9 percent. Growth in GNI (GDP at market prices plus net factor income) is slightly slower while growth in GNDI (GNI plus net current transfers) is much slower due to the decline in transfers (as grant aid declines). Growth in absorption (total domestic consumption and investment—that is, final domestic demand) is slower than GDP growth, at only 0.8 percent for the longer period. This is related to the balance of payment constraint and the loss in aid—over time, the economy is forced to drastically reduce its (goods and services) trade deficit from about 60 percent of GDP in 2009/10 to around 1 percent in 2025 in the BASE case, following the decline in grant aid. As a result exports have to grow much more rapidly than imports that even decline (7.6 percent a year compared with 0.9 percent a year), imposing the lower growth in domestic demand growth than in GDP. This scenario also results in a significant depreciation of the real exchange rate by around 1.9 percent per year, or a total depreciation of around 32 percent between 2010/11 and 2025/26. Note that the depreciation is especially strong in the transition phase resulting in an annual depreciation of 2.8 percent between 2010/11 and 2018/19. This effect, in particular, will somewhat boost exports and limit imports to the extent needed to maintain a balance in the external accounts. This may seem surprising as there is an expected appreciation pressure from the expansion of mining. However, the decline in aid is significantly larger, resulting in a net effect of a depreciating exchange rate, encouraging exports rather than challenging the competitiveness of nonmining exports. Among the components of domestic final demand (absorption), (on-budget) government consumption grows at 4.2 percent and investment at 3.7 percent, which is more rapidly than private consumption and investment. Private consumption and investment (domestic and FDI) grow at 4.2 percent and 2.9 percent, respectively, while donor consumption and investment declines by 13.3 percent and 8.8 percent, respectively. The assumed gradual increase in the household savings rate underpins private investment growth. Total (public and private) investment decreases with about 4 percent annually. 77 This means that for every 100 Afghanis worth of public investments total value added will increase by 14.2 the following year. 78 The projection treats the recurrent spending of government and donors as related O&M costs and expects the benefitting sectors to grow at the same rate as aggregate GDP. 118 Table A1.2 Real macro indicators by simulation (percentage annual growth) (2010/11–2025/26) 2010 BASE AGR– AIDALLOC AID— GOV– MIN+ GDP at factor cost 810.7 4.9 3.8 5.0 4.5 –0.5 5.8 Consumption Private 677.7 4.2 2.8 4.2 3.7 –0.5 4.4 Donors 348.9 –13.3 –13.2 –17.3 –13.3 –36.3 –13.4 Government 81.9 4.2 4.4 5.2 4.1 0.9 5.6 Fixed investment Private 53.4 2.9 1.6 3.1 2.6 –8.6 4.9 Donors 277.3 –8.8 –8.7 –12.9 –8.9 –40.9 –8.8 Government 63.7 3.7 3.2 5.9 3.4 –2.9 5.3 Exports 263.1 7.6 6.4 7.4 7.2 1.1 9.2 Imports 915.3 –0.9 –1.9 –1.0 –1.2 –6.9 –0.1 Absorption 1,502.9 0.8 –0.2 0.9 0.5 –4.3 1.4 Real exchange rate (index) 1.9 2.4 1.8 1.8 –0.4 1.7 (2010/11–2018/19) 2010 BASE AGR– AIDALLOC AID— GOV– MIN+ GDP at factor cost 810.7 5.9 4.9 6.1 5.2 0.7 7.0 Consumption Private 677.7 5.7 4.6 6.0 5.0 1.8 5.9 Donors 348.9 –14.1 –13.9 –21.3 –23.5 –56.7 –14.1 Government 81.9 1.6 2.0 7.4 1.6 –4.9 2.6 Fixed investment Private 53.4 3.5 2.4 4.3 2.8 –14.7 10.3 Donors 277.3 –10.2 –10.1 –17.7 –16.0 –63.3 –10.2 Government 63.7 1.3 1.2 9.8 1.0 –7.9 2.6 Exports 263.1 10.1 8.9 9.2 9.6 4.3 12.5 Imports 915.3 –2.2 –2.9 –2.8 –4.3 –10.7 –1.2 Absorption 1,502.9 0.1 –0.7 0.3 –1.4 –6.4 0.7 Real exchange rate (index) 2.8 3.4 2.3 2.9 1.8 2.8 Source: Authors’ calculations. Sectors will respond differently to the changes in economic fundamentals, not least due to due to relative price changes in favor of relatively tradable sectors driven by the depreciation of the exchange rate. The nonmining industry follows the economywide annual average for the 2011/12–2025/26 period (4.9 percent), while agriculture grows faster (6.4 percent) due to an assumed increase in investment in this sector. The service sector grows much slower than before, partly due to a reduction in donor consumption. In particular, donor security and nonsecurity services shrink, at annual rates of 35.1 and 4.2 percent respectively, while government security and nonsecurity services grow with a 3.9 and a 4.5 percent rate, respectively. However, even private services are affected by the decline in aid and cannot take advantage of the low exchange rate to the same extent as other more tradable sectors, resulting in a growth rate of only 2.9 percent. Mining, which starts from close to zero, grows at a very high rate of 60.1 percent; however, given its small initial size, it is more meaningful to assess the mining sector on the basis of its GDP shares (table A1.3). Opium growth is just above zero (given a fixed land area and slow productivity growth, both reflecting government strategies to constrain the sector). 119 Table A1.3 Average annual growth in value added, by activity and simulation, 2011/12–2025/26 (percent) BASE AGR– AIDALLOC AID— GOV– MIN+ Agriculture 6.4 3.9 6.2 5.9 1.6 6.3 Opium 0.1 0.1 0.1 0.1 0.0 0.1 Mining 60.1 60.1 60.1 60.1 1.1 68.7 Industry (nonmining) 4.9 4.3 5.0 4.4 –1.8 5.3 Services Private 3.0 1.9 3.1 2.5 –2.7 3.5 Government Security 3.9 4.2 4.2 3.9 1.7 5.2 Nonsecurity 4.5 4.7 6.4 4.3 –0.3 6.0 Donors Security –35.1 –35.0 –38.0 –35.1 –41.9 –35.1 Nonsecurity –4.2 –4.0 –8.6 –4.2 –30.6 –4.2 Total 4.9 3.8 5.0 4.5 –0.5 5.8 Source: Authors’ calculations. As shares of nominal GDP the major change in 2025/26 compared with 2010/11 is that mining increases from close to zero to almost 12 percent of GDP (table A1.4). Another major change is the decline of private services from 20 percent of GDP to only 13 percent. Together with more or less constant government services and a decline donor services (however, due to a high foreign wage share and imported goods, they never represented a large share of domestic GDP), overall services decreases from 30 to 22 percent of GDP. Agriculture increases its share from 28 to 33 percent, while the share of nonmining industries decreases from 28 to 24. Hence, we see a structural transformation with increased emphasize on mining but also agriculture, while especially services but also other industries decline in importance. Table A1.4 Shares of GDP at factor cost by activity in first report year and by simulation in final report year (percent) 2010 BASE AGR– AIDALLOC AID– GOV– MIN+ Agriculture 28.0 33.0 34.0 32.1 32.3 27.0 27.8 Opium 13.9 9.4 11.4 9.2 9.8 16.1 7.8 Mining 0.0 11.7 14.2 11.4 12.2 0.0 21.5 Industry (nonmining) 27.8 24.4 21.2 24.7 24.2 31.6 22.0 Services 30.3 21.6 19.2 22.6 21.5 25.3 20.8 Private 19.8 13.3 10.6 13.4 13.2 19.5 12.0 Government Security 5.5 5.4 5.6 5.7 5.4 4.1 5.8 Nonsecurity 2.7 2.6 2.7 3.5 2.6 1.7 2.8 Donors Security 1.4 0.0 0.0 0.0 0.0 0.0 0.0 Nonsecurity 0.9 0.2 0.2 0.1 0.2 0.0 0.2 Total 100 100 100 100 100 100 100 Source: Authors’ calculations. Considering sectoral growth rates and GDP shares, table A1.5 present the contribution to overall growth by sector. Mining is the main contributor with 2.4 percent of the 4.9 percent total growth (and 2.8 percent of the total annual 5.9 percent growth up until 2018/19), primarily due to very high sector growth. Mining is followed by agriculture with a 1.3 percent contribution, due to both decent growth and a high GDP share, closely followed by nonmining industry (0.9 percent). The nonmining industry, which includes construction, is negatively affected by the withdrawal of aid, but positively affected by the expansion of mining and its construction of ancillary infrastructure. Services are less important to 120 growth (contributing 0.4 percent) compared to the last decade, due to the withdrawal of aid and hence decline in the demand for private services (including transport) and in public services. Government service contribution to growth is negligible for the transition period in the BASE scenario, but becomes increasingly more important over time. Table A1.5 Decomposition of growth in GDP at factor cost by activity and simulation from first to final report year (percent) (2010/11–2025/26) BASE AGR– AIDALLOC AID— GOV– MIN+ Agriculture 1.3 0.7 1.3 1.1 0.2 0.9 Opium 0.0 0.0 0.0 0.0 0.0 0.0 Mining 2.4 2.4 2.4 2.4 0.0 3.8 Industry (nonmining) 0.9 0.6 0.9 0.7 –0.3 0.7 Services 0.4 0.2 0.4 0.3 –0.4 0.4 Private 0.3 0.2 0.4 0.3 –0.3 0.3 Government Security 0.1 0.1 0.2 0.1 0.0 0.2 Nonsecurity 0.1 0.1 0.1 0.1 0.0 0.1 Donors Security –0.2 –0.1 –0.2 –0.2 –0.1 –0.1 Nonsecurity 0.0 0.0 0.0 0.0 –0.1 0.0 Total 4.9 3.8 5.0 4.5 -0.5 5.8 (2010/11–2018/19) BASE AGR– AIDALLOC AID— GOV– MIN+ Agriculture 1.6 1.1 1.5 1.4 4.3 1.1 Opium 0.0 0.0 0.0 0.0 0.0 0.0 Mining 2.8 2.8 2.8 2.8 0.0 4.8 Industry (nonmining) 1.0 0.8 1.1 0.8 0.1 0.8 Services 0.3 0.2 0.7 0.0 –3.8 0.3 Private 0.4 0.2 0.4 0.2 –1.2 0.3 Government Security 0.0 0.0 0.3 0.0 –0.5 0.0 Nonsecurity 0.0 0.0 0.2 0.0 –0.3 0.0 Donors Security –0.1 –0.1 –0.1 –0.2 –1.1 –0.1 Nonsecurity 0.0 0.0 0.0 0.0 –0.6 0.0 Total 5.9 4.9 6.1 5.2 0.7 7.0 Source: Authors’ calculations. Constrained agricultural growth, AGR- Historically agriculture has been a large source of weather dependent output volatility. While the BASE scenario assumes that some more investment in the agriculture sector remedies some of this volatility (e.g. though investment in irrigation technology) which results in growth at 6.4 annually on average, the AGR- scenario assumes the average historic growth rates for the sector (3.9 percent). With the lower growth in the agricultural sector, overall growth declines from 4.9 percent in the BASE scenario to 3.8 percent annually on average between 2010 and 2025, and from 5.9 to 4.9 in the transition period. Absorption declines by an average rate of 0.2 percent due to a decline in domestic demand. The main difference to the BASE scenario is the slow-down in growth of private consumption growth from 4.2 percent to 2.8, and in growth of private investment from 2.9 to 1.6 percent. Export growth also declines 121 compared to the BASE (from 7.6 to 6.4 percent) despite the higher rate of depreciation of the exchange rate (grows with 2.4 percent rather than 1.9 on average) resulting not least from the lower GDP growth. Alternative aid dynamics Higher share of aid on the government budget, AIDALLOC The government budget is highly dependent on aid. In 2010/11, the share of on-budget aid in total aid was 12 percent, which increased to almost 17 percent in the BASE case. Within the next years, the mutual goal of the government and donors is to move a substantial part of aid to the government (on- budget) from donors (external budget).79 The objective of the AIDALLOC scenario is to demonstrate the economic effects of a major shift in aid modality. Departing from the BASE scenario, the AIDALLOC scenario reallocates more aid from the external donor budget to the government budget without any further changes in other variables. It is assumed that the total amount of aid (in FCU) declines with the same gradual speed than in the BASE case and that the spending shares remain unchanged for total public security vs. total nonsecurity and total consumption vs. total investment. Under AIDALLOC, the shares of on and external budget change drastically in 2011/12 and gradually shift during the rest of the simulation period, leading to shares of 59 percent on- budget and 41 percent external budget in 2025/26. This simulation is expected to have an impact primarily due to differences in local content (domestic demand share) between the government and donor budget, possibly resulting in a higher domestic impact when aid is on-budget. Table A1.6 presents the resulting domestic demand share in MAMS given assumption about import of goods and labor for the specific activity and related investments (import shares of intermediate value added is included). Table A1.6 Domestic demand assumptions of different types of aid Type of aid Domestic impact Government-managed Nonsecurity 69 Security 96 Donor-managed Nonsecurity 25 Security 14 Source: The World Bank Note: These calculations include adjustment due to the use of foreign labor, directly imported goods, imported content in sectors used as intermediate inputs, and any capital remitted. However, the final economic effects of the aid reallocation are relatively minor compared to the BASE scenario. Average annual growth in GDP increases from 4.9 to 5.0 percent, for the period 2011/12– 2025/26 reflected in absorption growth that increases from 0.8 to 0.9 percent. Looking over time, GDP growth spikes in 2011/12 as a result of the sudden shift in resources, but revert to lower levels of growth during the rest of the period. Hence, the average annual growth rate for the transition period 2011/12–2018/19 was as high as 6.1 percent. Among absorption (consumption and investment), growth accelerates for the government and declines further for donors as a consequence of the reallocation, but in addition private investments growth increases from 2.9 to 3.1 percent. Export growth decreases 79 The agreed objective at the Kabul conference in 2011 was to shift 50 percent of donor funding on-budget. 122 slightly compared to the BASE mainly due to the lower depreciation of the real exchange rate, resulting from the higher local content of on-budget aid and the higher GDP growth. Poverty and unemployment both stay more or less the same as the BASE. Rapid decline in both security and nonsecurity aid, AID- Under this scenario, donor-managed (external budget) grant aid (defined in foreign currency) declines more rapidly, reaching 15 percent of the 2010/11 levels already by 2016/17 (for the BASE this took until 2025/26) and staying at that level until 2025/26 (figure A1.4). At the same time, starting from 2016, the donors switch almost all of their aid to nonsecurity. Figure A1.4 On-budget aid and external aid developments in the AID- scenario (rapid decline), 2009/10–2025/26 (2009/10 USUS$ billion) 9 8 Donor, security 7 6 2009/10 $ billion 5 Donor, nonsecurity 4 3 2 Government, nonsecurity 1 Government, security 0 2009/10 2011/12 2013/14 2015/16 2017/18 2019/20 2021/22 2023/24 2025/26 Source: Authors’ calculations. The economic impact from a rapid decline in external aid is significant for several key economic variables. This is especially true during the transition period (2010–18) but also for the final annual average numbers in 2025/26. Annual growth in GDP declines to 4.5 percent (growth for the BASE was 4.9 percent), and growth in absorption (domestic demand) declines from 0.8 to as little as 0.5 percent. 80 However, the drop in annual average GDP growth rate was more pronounced for the transition period (from 5.9 to 5.2 percent). Among domestic demands, annual growth rates for private consumption declines with 0.5 percentage points compared to the BASE and private investment by 0.3. The decline is weaker for government consumption due to the importance of wages in government services and 80 The decline in GDP growth is due to losses both in factor employment and TFP, with a more important role played by the latter. 123 slower wage growth in response to the slowdown in GDP growth. The aid loss for donors is the same for AID- and the BASE by 2025/26, but it was more drastic in the beginning for AID- resulting in a substantially smaller accumulated donor over the whole period. Exports growth declines by 0.4 percentage points, due to lower growth in general but also somewhat less depreciation. Import growth declines even further, at a rate of 1.2 percent instead of 0.9, due to the high import content in external aid. The exchange rate index grows by 1.8 percent instead of 1.9 in the BASE, resulting in an appreciation compared to the BASE. The impact on the real exchange rate varies over time; it is more strongly depreciated than for the BASE scenario up to 2015/16 when it turns more appreciated; the main factors that pull the exchange rate in different directions are (a) appreciation due to weaker GDP growth and lower growth for donor-run consumption and investment, which is highly import-intensive; and (b) depreciation due to less foreign grant aid. Not unexpectedly, private service growth declines compared to the BASE (by 0.6 percentage points) as that sector so far has gained from the substantial inflow of aid. Nonmining industries are also affected (presumably due to the lower growth in construction following the decline in aid), as well as agriculture growth, which also declines. Nonmining industry and agriculture growth is affected by the lower accumulated public capital stock by the early and rapid decline in aid. Mining is less affected by aid. Alternative developments of the mining sector Additional mining expansion, MIN+ In addition to the well-known Aynak and Hajigak mines, eleven other highly prospective mines have been identified. IFC (2011b) has estimated the expected production value, operational costs, and investment needs (taking into account the public character of many investments that could be shared among several mines) for a scenario that, in addition to Aynak and Hajigak, also includes these 11 additional mines. On the basis of this database, we simulate the impact of such a stronger mining expansion. Figure A1.5 shows the resulting mining value added magnitude of this scenario compared to the BASE. 124 Figure A1.5 Additional mining value added in BASE and MIN+ (constant US$ million) 6,000 MIN+ 5,000 4,000 Constant $ million 3,000 BASE 2,000 1,000 0 2010/11 2012/13 2014/15 2016/17 2018/19 2020/21 2022/23 2024/25 Source: Author’s calculations, based on initial data from IFC (2011b). Naturally, the result shows the additional exploitation of mines would significantly affect growth and other key variables. Apart from the mining value added growth itself, overall growth is encouraged by the higher public capital stock resulting from the additional government revenue and the fact that the part of mining investments that will be available to the rest of the economy (including roads and power plants; some 20 percent of total mining investments). According to the simulation results, GDP growth would increase from 4.9 percent in the BASE case (which only includes Aynak and Hajigak are developed, to 5.8 percent with the additional 11 mines. For the transition period the average annual growth would increase from 5.9 to as much as 7.0 percent. Donor consumption and investments are similar to the BASE, but growth in government consumption and expenditure increases from 4.2 to 5.6 and 3.7 to 5.3, respectively, due to stronger growth in fiscal resources. Household consumption and especially household investment also increases; growth in private investments is 4.9 percent on average instead of 2.9 in the BASE case with less mining exploitation. Absorption consequently grows by 1.4 percent annually instead of 0.9 in the BASE. The real exchange rate depreciation is lower due to the increased inflow of foreign currency related to mining—coming both from increased exports and increased FDI—even though this effect is mitigated by remitted profits and the relatively low domestic content in mining intermediate demands and investment. Exports grow by as much as 9.2 percent per year, which is 1.6 percentage points higher than in the BASE. In terms of sector effects, the contribution of mining to growth increases from 2.4 to 3.8 percent. Behind this is a mining specific annual sector growth that is as 69 percent. The resulting mining share in 2025/26 is 21.5 percent of GDP, which is similar or slightly higher than the shares for each of nonmining 125 industry and private services. Agriculture and opium grow at rates similar to those of the BASE scenario, while the growth rates of nonmining industries, private services both increase by 0.4 percentage points and government services expand by an additional 1.4 percentage points. Deteriorated security and governance situation, GOV- GOV- presents a scenario where the security and governance situation is rapidly worsening with substantially lower levels of aid, effects on savings and investments and no mining developments. The deteriorated security situation is assumed to lower saving rates to 2/3 of the 2010/11 levels in 2013 and then to 1/3 by 2014, double the depreciation rate of both public and private capital, and reduce productivity growth to almost zero. Total aid is drastically dropped to 1/3 of the BASE value of total aid in 2013 and further to about 1 percent in 2016, resulting in a final aid as a percentage of GDP in 2025/26 of 0.4 percent. In the GOV- case, annual average GDP growth is even negative—that is, GDP decreases during 2010/11– 2025/26. GDP growth is -0.5 percent compared to 4.9 in the BASE case. In the shorter transition period the annual drop in services is more severe but growth is somewhat better (0.7 percent) due to continued agriculture production. However, even agriculture growth is declining significantly in the long run.81 Domestic demand (absorption) declines even more (-4.3 percent) mainly due to the drastic decline in donor consumption and investment that are decreased by 36 and 41 percent respectively. However, government consumption only grow by 0.9 percent and government investments declines with 2.9 percent, partly due to the decline in aid and partly due to the low GDP growth and lack of mining development, limiting tax revenue (13.6 percent of GDP compared to 15.0 percent in the BASE). Private consumption and investment decline with 0.5 and 8.6 percent respectively. Consequently, poverty is strongly affected resulting in a poverty rate in 2025/26 of 63 percent compared to 36 in 2010/11. Un/under employment increases from 28 to 35 percent. The real exchange rate is appreciating with an average rate of 0.4 percent despite the drastic reduction in aid and slower growth. This is probably due to the reduction in imports resulting from the lower shares of government and donor activities, both heavy in imports; especially their investments which are strongly affected in this scenario. Export growth declines to as little as 1.1 percent per year as a result of the lack of mining developments and the absence of a depreciated exchange rate that could encourage exports from other sectors. At the sector level we see the expected sharp decline in donor services but also in private services that declines by an average growth of 2.7 percent. Also nonmining industry declines over the period while agriculture, mining and government services grow but at a limited rate. As final shares of GDP in 2025/26 there are marginal changes compared with 2010/11, with the most noticeable in the increase in opium, from 13.9 of GDP to 16.1 percent, as result of households seeking alternative livelihoods and a decrease in services, from 30.3 of GDP to 25.3 percent. 81 To get a sense of the mechanism leading up to this negative growth we ran an additional scenario similar to the BASE but with the substantially lower aid levels and negative effects on the saving rate, the depreciation rate and general productivity from security, representing a case where the security situation is deteriorated as in GOV- but mining is still being developed. This scenario results in an average growth rate of 1.1 percent, i.e. the lack of aid and the deterioration in security and its assumed effects on savings, the capital stock and productivity accounts for most of the decline in growth in the GOV- scenario. 126 Annex: Tables for 2010–18 Table A1.7 2010 BASE AGR– AIDALLOC AID– MIN+ MIN– GOV– INV CONS GDP at factor cost 810.7 5.9 4.9 6.1 5.2 7.0 5.1 0.7 6.0 5.6 Consumption Private 677.7 5.7 4.6 6.0 5.0 5.9 5.5 1.8 5.8 5.4 Donors 348.9 –14.1 –13.9 –21.3 –23.5 –14.1 –14.0 –56.7 –15.4 –11.9 Government 81.9 1.6 2.0 7.4 1.6 2.6 –0.6 –4.9 0.2 7.4 Fixed investment Private 53.4 3.5 2.4 4.3 2.8 10.3 3.1 –14.7 3.6 3.5 Donors 277.3 –10.2 –10.1 –17.7 –16.0 –10.2 –10.1 –63.3 –9.2 –12.7 Government 63.7 1.3 1.2 9.8 1.0 2.6 –0.9 –7.9 3.0 –15.2 Exports 263.1 10.1 8.9 9.2 9.6 12.5 4.3 4.3 10.3 9.0 Imports 915.3 –2.2 –2.9 –2.8 –4.3 –1.2 –10.7 –10.7 –2.0 –2.9 Absorption 1,502.9 0.1 –0.7 0.3 –1.4 0.7 –0.3 –6.4 0.1 –0.1 Real exchange rate (index) 2.8 3.4 2.3 2.9 2.8 3.3 1.8 2.9 2.6 Total factor employment (index) 3.9 3.8 3.9 3.5 4.2 3.6 1.6 4.0 3.6 Unemployment rate (percent) 27.7 27.2 27.9 27.1 27.7 27.0 27.2 30.4 27.1 27.4 Headcount poverty rate (percent) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 127 Table A1.8 BASE AGR– AIDALLOC AID— MIN+ MIN– GOV– INV CONS Agriculture 8.6 6.3 7.7 8.0 8.5 9.0 5.1 8.9 7.5 Opium 0.2 0.2 0.2 0.2 0.2 0.2 0.1 0.2 0.2 Mining 122.8 122.8 122.7 122.8 145.5 1.0 1.1 122.8 122.8 Industry (nonmining) 6.1 5.7 6.4 5.3 6.7 6.2 0.1 6.4 4.9 Services Private 3.4 2.6 3.6 2.2 3.8 3.2 –2.4 3.4 3.4 Government Security 1.1 1.6 6.6 1.3 2.1 –1.1 –4.5 1.2 0.9 Nonsecurity 2.2 2.6 8.6 2.0 3.3 0.0 –5.5 –1.3 13.8 Donors Security –17.7 –17.5 –24.6 –55.5 –17.7 –17.6 –58.5 –17.7 –17.7 Nonsecurity –6.3 –6.1 –14.2 –7.7 –6.3 –6.2 –52.7 –9.6 –1.3 Total 5.9 4.9 6.1 5.2 7.0 5.1 0.7 6.0 5.6 Table A1.9 2010 BASE AGR– AIDALLOC AID– MIN+ MIN– GOV– INV CONS Agriculture 28.0 32.6 33.5 30.4 32.9 29.6 37.0 34.8 33.0 31.2 Opium 13.9 11.5 12.6 10.7 12.1 10.3 12.9 16.1 11.4 11.4 Mining 0.0 7.3 8.1 6.9 7.7 14.4 0.0 0.0 7.3 7.3 Industry (nonmining) 27.8 26.4 24.7 26.6 25.9 24.7 27.9 28.3 26.8 24.5 Services 30.3 22.2 21.1 25.4 21.3 21.0 22.3 20.8 21.5 25.6 Private 19.8 15.3 13.9 15.3 14.6 14.2 15.9 16.5 15.2 15.7 Government Security 5.5 4.3 4.4 6.4 4.3 4.2 3.9 2.9 4.3 4.2 Nonsecurity 2.7 2.1 2.2 3.4 2.1 2.1 1.9 1.3 1.6 5.0 Donors Security 1.4 0.2 0.2 0.1 0.0 0.2 0.2 0.0 0.2 0.2 Nonsecurity 0.9 0.3 0.3 0.2 0.3 0.3 0.4 0.0 0.3 0.5 Total 100 100 100 100 100 100 100 100 100 100 128 Table A1.10 BASE AGR– AIDALLOC AID– MIN+ MIN– GOV– INV CONS Agriculture 1.6 1.1 1.5 1.4 1.1 3.0 4.3 1.7 1.4 Opium 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Mining 2.8 2.8 2.8 2.8 4.8 0.0 0.0 2.9 2.8 Industry (nonmining) 1.0 0.8 1.1 0.8 0.8 1.7 0.1 1.1 0.8 Services 0.3 0.2 0.7 0.0 0.3 0.3 –3.8 0.3 0.6 Private 0.37 0.24 0.41 0.22 0.29 0.58 –1.18 0.38 0.37 Government Security 0.03 0.04 0.26 0.04 0.05 –0.05 –0.51 0.04 0.03 Nonsecurity 0.03 0.03 0.17 0.03 0.04 0.00 –0.30 –0.02 0.33 Donors Security –0.09 –0.08 –0.12 –0.24 –0.07 –0.15 –1.13 –0.09 –0.09 Nonsecurity –0.02 –0.02 –0.05 –0.03 –0.02 –0.04 –0.64 –0.04 –0.01 Total 5.9 4.9 6.1 5.2 7.0 5.1 0.7 6.0 5.6 129 Table A1.11 2010 BASE AGR– AIDALLOC AID– MIN+ MIN– GOV– INV CONS Government Revenues Domestic revenues 8.3 10.7 10.9 10.4 10.7 11.0 9.1 8.8 10.7 10.4 Grants 10.0 2.9 3.2 12.2 3.1 2.3 3.3 0.4 2.9 2.9 Foreign borrowing 0.6 0.8 0.9 0.8 0.9 0.8 0.9 1.2 0.8 0.8 Expenditures Security 5.8 4.5 4.7 7.8 4.6 4.5 4.1 3.2 4.5 4.4 Nonsecurity 11.3 8.8 9.4 14.4 9.0 8.7 8.0 6.2 8.8 8.6 Household transfers 1.7 1.2 1.2 1.1 1.1 1.1 1.2 1.0 1.2 1.2 External donor budget Revenues Grants 73.6 20.0 22.0 9.4 10.5 18.1 22.3 0.1 20.0 20.0 Expenditures Security 45.6 7.3 8.0 3.4 0.1 6.6 8.1 0.0 7.3 7.3 Nonsecurity 28.0 12.7 13.9 6.0 10.4 11.5 14.2 0.0 12.7 12.7 Total aid 83.6 23.0 25.2 21.6 13.6 20.7 25.6 0.5 22.9 23.0 On-budget aid 11.9 12.8 12.8 56.4 22.7 12.8 12.8 86.0 12.8 12.8 (percentage of total aid) Security aid 58.3 35.8 35.8 34.7 7.9 35.8 35.8 33.3 35.8 35.8 (percentage of total aid) 130 References Bourguinon, Francois, Carolina Diaz-Bonilla, and Hans Lofgren, 2008. “Aid, Service delivery and the Millennium Development Goals in an Economy wide Framework,” pp. 283-315 in Francois Bourguinon, Maurizio Bussolo, and Luiz A. Pereira da Silva, eds. The Impact of Macroeconomic Policies on Poverty and Income Distribution: Macro-Micro Evaluation Techniques and Tools. Washington, D.C., World Bank. Lofgren, Hans, Martin Cicowiez, and Carolina Diaz-Bonilla. 2012. “MAMS – A Computable General Equilibrium Model for Developing Country Strategy Analysis.” Chapter prepared for Handbook of CGE Modeling, editors, Dale Jorgenson and Petr Dixon. Forthcoming. 131 Appendix 2 Fiscal Technical Background Macro assumptions This fiscal model is based on the assumption that the security situation remains constant. The basic macroeconomic assumptions of the model are the same as in the growth baseline scenario —real gross domestic product (GDP) growth averaging 5 percent, stable Inflation at 5 percent, and exchange rate depreciation from 47 to 54 Af/US$ by 2021/22. Mining revenue Estimates for mining revenue are based on the Aynak and Hajigak projects alone. The estimates take into account the construction and operating phases of each mine and the profits and taxes calculated for the mine itself and not for ancillary infrastructure. During the construction period, 2011–15, it is assumed that no significant taxes are paid by the mining companies except bonus payments. But in the operating phase, 2016–22, the capital cost of the initial investment are amortized and it is necessary to determine profits and profit-based income taxes paid by each mine. The output of the mine is multiplied by unit prices to get gross revenue. For copper, the average London Metal Exchange price over first half of 2011 of US$ 8,875 per tonne is used.82 For iron ore, the average free on board contract price to Europe of US$ 160 per tonne is used. For Aynak, costs are based partly on information provided by the company and mid-range operating costs globally. The estimated cost of pellet production at Hajigak is US$ 75 per tonne, which is based on mid-range, global production costs. Profits and profit taxes are then calculated allowing for amortization, which feeds into government revenue and expenditure, as well as corresponding multiplier impacts. The development included in this study for Hajigak follows the general design of the Aynak development, and is assumed to produce 10 million tons of iron ore pellets.83 The development includes a 300 megawatt coal power plant, fueled by coal from a mine operated by the mine owner through a concession from the government; 50 percent of the power is to be provided to the community/region at cost. (It is assumed that the developer will also be required to build a railroad, though this is not included in the analysis.) 82 It could be argued that for copper the current 27 month futures price (the farthest available) of US$ 9,210/tonne is the best estimator, but given that copper prices have been much higher than their secular trend in recent months, the average price over the past year was used. The similar logic was used in the case of iron ore where there is no large futures market but the most recent price (June 2011) is US$ 172/tonne. 83 This is a modest assumption considering the size of the deposits identified at Hajigak. The results from the analyses below can be scaled, within reason, to accommodate a larger development. 132 Customs revenue Recorded imports data from fiscal year 2005/6–2010/11 (1384–1389) from the Afghan Customs Department was analyzed by the Harmonized System code to determine the value of imports across various sectors. Imports were categorized in six main groups—food; petroleum products; construction materials; industry; vehicles, parts, tires, and machinery; and consumables and consumer goods.84 Total customs collection for the period across various sectors was also determined as custom duties, business receipt taxes, and other charges. Correlations between historical levels of aid and imports were tested through simple regression models. Except for the food and industry groups, the value of imports in the remaining groups was found to be significantly correlated with total aid.85 A nonsignificant correlation between food-related imports with aid signifies that such imports are, as intuitively expected, independent of aid. Similarly for the industry group, no correlation with aid implies a weak industry sector with negligible aid-related growth. This conclusion is supported by other studies, where the contribution of Afghan private sector investment in industry remains weak with a negligible contribution to job creation and the economy (World Bank and DfID 2008). Other major underlying assumptions for the above analyses are that with the withdrawal of international security forces in 1393, and reduction in aid—the security situation does not worsen, systems overall remain stable, border controls do not get any worse, and protectionist tariffs are not introduced. All or some of these events could possibly take place. Value-added tax Value added tax estimates are in line with the IMF baseline scenario for Afghanistan (IDA and IMF 2011). Security wage bill The estimates of the security wage bill take into account the costs for Afghan National Army and Afghan National Police troops for both salary and food allowances. This leads to a security wage bill that grows from 8 percent to 11 percent of GDP in 1389–1394 and stays more or less constant. Civil service wage bill The estimates for the civil service wage bill are based on costing for the pay and grading reform informed by Civil Service Commission figures and progress to date (1389). It assumes an increase from 5 percent to 7 percent of GDP as a result of concluding the pay and grading efforts, and assumes an extra 1 percent of GDP that is driven by other line ministry pressures of increasing the tashkeel, particularly as a result of increased hiring for new teachers. 84 Afghan Customs Department/Tariff Research Statistics Unit data from 2005/6–2009/10 (1384 through 1388) for imports and custom revenues, and data from the Annual Review on Tariff, External Trade and Revenue (1389). 85 Considering the small sample size, 1386 data were clearly an outlier and had to be excluded from the regression analysis. 133 Security operation and maintenance costs The International Security Assistance Force estimates that US$ 3.5 billion would be required a year for operation and maintenance (O&M) of military infrastructure and equipment, excluding salaries, for existing and under construction assets. This figure considers the requirements to operate and maintain the Force’s reported US$ 11.4 billion investment in security infrastructure from 2006 to 2014 and would cover utility bills, minor repairs, and everyday operations of the military bases and police stations. In addition, it includes the costs to the Afghan National Security Forces of existing and planned new equipment related to fuel, spare parts, minor repairs, contractor services, and travel allowances. Civilian O&M costs This analysis estimates that US$ 1.3 billion would be required annually to operate and maintain the stock of civilian assets. The figure is based on the technical O&M needs to support assets listed in a constructed registry; it accounts for intrasectoral and intersectoral variations. The calculation of civilian O&M needs is based on central and provincial research commissioned by the World Bank for this report and required the construction of a comprehensive public sector asset registry to determine the bulk of the existing and soon-to-be handed over assets in need of O&M in the coming years. Fieldwork was conducted in the provinces of Uruzgan and Bamyan. For more information on O&M needs and estimation, see Altai Consulting (2011). References Altai Consulting. 2011. Assessing O&M of Key Assets in Afghan Provinces. Kabul: World Bank. IDA/IMF, 2011,Islamic Republic of Afghanistan, Joint Bank/Fund Debt Sustainability Analysis, prepared by staffs at the IDA and IMF. World Bank and DFID (UK Department for International Development). 2008. “The Afghanistan Investment Climate Assessment”. 134 Appendix 3 Range of civil servants’ remuneration mechanisms 2011 Payment Description Rates Comments modality Current pay Paid to civil servants who have US$ 56–US$ 206 The number of staff on these modalities is scale not gone through pay and a month for 12 related to P&G. As P&G numbers increase, grading (P&G) reform grades these decrease. To be phased out by the time P&G reform is completely rolled out by end- Priority Interim additional allowances US$ 80–US$ 235 2014 reform and for staff in line a month for six restructuring ministries/agencies who have grades undertaken reform of structure as part of their job description Superscale To remunerate the top cadres US$ 300 to US$ Payment of superscale salaries is of civil servants who have 1,500 a month discretionary, but to attract and retain undergone P&G reform at a for grades 1–4 qualified staff, top-spending ministries hire higher rate more civil servants on these salaries. The large ministries pay superscale salaries to 30– 40 staff on average Presidential Awards made to individuals Variable up to Salary levels are largely assigned through awards through a presidential order, US$ 6,000 a negotiations between the Independent mainly to deputy ministers, month Administration Reform and Civil Service chiefs of staff, and general Commission, Ministry of Finance, and the directors Office of the President Cadre Job specific allowances for Hazard pay for MoPH pays allowances for high-risk jobs such allowances attorneys, judges, engineers, insecure as radiologists and laboratory workers. and teachers employed by the provinces Ministry of Higher Education has special Ministry of Higher Education, allowances for lecturers. Ministry of Justice specialists with the Ministry of pays special allowances for senior judges and Public Health (MoPH) lawyers. The precise modality by which these cadres are paid is not yet fully defined and the civil servants law is being rewritten to include these benefits 135 Payment Description Rates Comments modality P&G reform Implementation of new P&G US$ 116–US$ Total number of positions regraded to date is structure throughout the 714 a month for around 303,000. Around 160,000 of these government. Merit-based 8 grades have completed the process through priority reform and regrading appointment to receive new salary restructuring employees transitioned automatically. This is based on the Civil Service Law Management Contract civil servants filling US$ 1,000–US$ Up to 142 total appointments will be made; Capacity grade 1–4 line positions 7,500 a month 80 are now in place Program covering core functions for 2– 3 years. Merit-based recruitment Salary top-ups Ministry-specific top-ups to Ministry and Revenue-generating ministries can use up to civil servants. Performance- or donor specific 5 percent of their revenue to provide top-ups incentive-based salaries, food to their staff. Guidelines on how revenue is allowances, per diems, and used for salary payment vary by ministry transport, hardship, and leave allowances 136