70851 v2 Afghanistan in  Transition: Looking  Beyond 2014  Volume 2: Main Report            May 2012    xxii  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  Mediumâ€?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 underâ€?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)  2  3 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 onâ€?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 onâ€?budget aid and external aid developments in the AIDâ€? 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 lowâ€?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 inâ€?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 standâ€?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 Elsewhereâ€?; Claudia Nassif: chapter 2 “The Economic Impacts of Transitionâ€?; Camilo Gomez Osorio:  chapter 3 “Managing the Fiscal Challengeâ€?; Richard Hogg: chapter 4 “Building Government Capacityâ€?;  Andrew Beath and William Byrd: chapter 5 “Delivering Services and Maintaining Infrastructure.â€? 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 Coâ€?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 governanceâ€?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 enoughâ€? 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  conflictâ€?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 midâ€?18th century, formed as  an  expansionist  Pashtunâ€?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 invasionsâ€? and  started  to  build  basic  institutions  of  a  central  state.  After  several  decades  of  semiâ€?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  nonâ€?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 preâ€?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 externallyâ€?driven measures  of  the  Sovietâ€?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 subsistenceâ€? 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, eliteâ€?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  ï‚· Removing Babrak Karmal, seen as an ineffective head of state and Afghan Communist Party leader,  in 1986, and replacing him with Najibullah.   ï‚· Reversing  the  Communist  reform  agenda  and  switching  to  a  National  Reconciliation  Policy,  that  called for a ceaseâ€?fire with mujahideen forces, talks, a transitional government, and elections.  ï‚· 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.  ï‚· 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 “guaranteedâ€? by the Soviet  Union  and  the  United  States—but  the  accords  neither  included,  nor  were  signed  by,  mujahideen  representatives.  ï‚· 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 faceâ€?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 coâ€?opted a range of tribal leaders and negotiated deals for local  control,  which  often  took  the  form  of  governmentâ€?backed  militias  paid  with  Sovietâ€?provided  aid  including  weapons,  food,  and  fuel.  This  approach  worked  in  the  short  term  but  carried  serious  risks,  since those coâ€?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:3                                                               3  See Fotini (2011) and Byrd (2012).  3  ï‚· Effective  Afghan  leadership  was  critical  in  the  Soviet  withdrawal  and  for  the  survival  of  the  postâ€? 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.  ï‚· The withdrawal of foreign military forces removed a major justification for mujahideen opposition,  allowing the government to exploit divisions among them.   ï‚· Total reliance on Afghan security forces after foreign military withdrawal was unrealistic, especially  without a meaningful peace agreement or at least a durable ceaseâ€?fire. The postâ€?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.  ï‚· 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.  ï‚· Durable peace was impossible without the buyâ€?in of key regional players, notably Pakistan. (Pakistan  was a signatory to the Geneva Accords, but the mujahideen groups based in Pakistan were not.)  ï‚· The  Geneva  Accords  did  little  more  than  provide  a  faceâ€?saving  cover  for  the  Soviet  military  withdrawal, and its shortcomings were evident with the failure to bring the mujahideen to the peace  table.   ï‚· The  postâ€?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:  ï‚· 2001–05:  optimism  for  the  prospects  of  building  a  modern,  developmental,  and  democratic  state  based on the Bonn Agreement of December 2001.  ï‚· 2005–09: Taliban resurgence and “militarizationâ€? of aid, which increasingly became a tool for shortâ€? term stabilization interventions in conflictâ€?affected areas.  ï‚· 2009–11: U.S. military and civilian surge and a dramatic increase in the U.S. aid budget.   ï‚· 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  postâ€?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  postâ€?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 clustersâ€? (weak, middle, and strong).4 It  also summarizes the current gap with these strong performers, indicating how far Afghanistan needs to  go to become something like a typically successful postâ€?conflict, lowâ€?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  Underâ€?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                                                               4  These comparisons are based on an “event studyâ€? 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 postâ€? 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 perceptionsâ€?based governanceâ€?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  postâ€? 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        Guineaâ€?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  OECDâ€?DAC data for Afghanistan, round black points are World Bank staff estimates.  9  Data on technical assistance (according to the OECD definition of “technical cooperationâ€?) 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 postâ€?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 loyaltyâ€? and provide security and political stability.   The key economic question for Afghanistan’s transition is whether large reductions in aid to other postâ€? conflict or conflictâ€?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  lowâ€?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 15â€?year transition  period. Yet real per capita GDP maintained average annual growth of over 4 percent. In Guineaâ€?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 15â€?year transition period.   In  Bosnia  and  Herzegovina,  a  middleâ€?income  postâ€?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 postâ€?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  middleâ€?income  country.  After  independence  in  Timorâ€?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 postâ€?conflict  period, when other trends—in particular the typical postâ€?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  yearâ€?toâ€?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 shortâ€?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  oneâ€?off  activities  such  as  rebuilding  infrastructure,  returning  refugees,  and  other  shortâ€?term  quickâ€?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 postâ€? conflict  economic  boom  that  fueled  growth  in  the  initial  postâ€?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 proâ€?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  clanâ€?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  allâ€?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. Clanâ€?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 pattern of “governance without governmentâ€?;  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:  ï‚· “Divide and ruleâ€? 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 powerâ€?holders to exploit.  ï‚· 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.  ï‚· Abruptly stopping foreign aid, especially during a conflict, can expose the weakness of a highly aidâ€?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).   ï‚· The emergence and entrenchment of a “war economyâ€? can fuel the persistence of conflict, as various actors— even if originally motivated by grievances or political considerations—benefit from the spoils.  ï‚· 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.  ï‚· Geopolitical  significance  (earlier  in  the  Cold  War  or  more  recently  in  the  U.S.â€?led  “war  on  terrorâ€?)  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  postâ€?conflict  transition,  as  with  underâ€?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 Subâ€?Saharan  Africa  affected  by  high  rates  of  HIV/AIDS.5  Internationally,  life  expectancy  increased  significantly  on  average for strong performers and middle performers, but stagnated or declined for weak performers.                                                               5  The 2010 Afghanistan Mortality Survey shows dramatic improvements in life expectancy and underâ€?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 timeâ€?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 underâ€?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 Underâ€?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,  underâ€?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 underâ€?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 betterâ€?performing comparator countries for  underâ€?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  perceptionsâ€?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 â€?0.5 â€?1.0 â€?1.5 â€?2.0 â€?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 â€?0.5 â€?1.0 â€?1.5 â€?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 “inclusiveâ€? enoughâ€?  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 â€?0.5 â€?1.0 â€?1.5 â€?2.0 â€?2.5 â€?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).  Perceptionsâ€?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  clearâ€?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  patronâ€?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  powerâ€?holders  rather  than  consolidation  of  political  parties.6   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  shortâ€?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.                                                                6  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  postâ€?2001  period  ushered  in  major  changes  but  also  has  strong  elements  of  continuity  with  the  past.  Critical  aspects  of  Afghanistan’s political economy are:  ï‚· 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.  ï‚· Considerable  insecurity  in  many  parts  of  the  country,  including  an  active  insurgency  but  also  local  conflicts, criminality, and predation by criminal groups.  ï‚· Ethnic  and  factional  fragmentation  and  associated  tensions,  and  a  lack  of  solid  and  durable  coalitions of common interest.  ï‚· 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  shortâ€?termâ€?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  shortâ€?term  risks  while  striving  to  exploit  the  opportunity  to  promote  more  positive  political  economy  outcomes  over  the  medium  term.  Possible  actions include the following:  ï‚· 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.  ï‚· Improving  aid  effectiveness  by  channeling  more  aid  through  the  budget:  this  may  help  reduce  the  political economy distortions that fragmented and offâ€?budget aid inadvertently encouraged.  18  ï‚· 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.  ï‚· Improving  the  process  for  government  appointments:  even  if  a  fully  meritâ€?based  system  will  take  time  to  take  hold,  the  appointment  process  should  be  more  inclusive  and  evenly  spread  across  different groups and factions.  ï‚· 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.   ï‚· 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  runâ€?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  “normalâ€?  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  shortâ€?termism,  factionalized  politics,  widespread  patronage,  and  corruption  and  more  consistently  pursue  a  mediumâ€?  to  longer  term  national  agenda.  This will require political consolidation and a broadâ€?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.â€?  Background  Paper, World Bank, Washington DC   Byrd, William, Gary Milante, and Kenneth Anye. 2012. “Afghanistan Transition Study: Quantitative and  Indicatorâ€?Wise International Comparisons.â€? Background Paper, World Bank, Washington DC.  De  Waal,  Alex.  2009.  “Mission  Without  End?  Peacekeeping  in  the  African  Political  Marketplace.â€?  International Affairs 85 (1): 99–113.  Fotini,  Christi.  2011.  “Lessons  from  the  Soviet  Withdrawal  for  the  Presentâ€?day  Afghan  Transition.â€?  Background paper, World Bank, Washington DC.  Keefer,  Philip.  2008.  “Insurgency  and  Credible  Commitment  in  Autocracies  and  Democracies.â€?  World  Bank Economic Review 22 (1): 33–61.  ———. 2010. “Why Follow the Leader? Collective Action, Credible Commitment and Conflict.â€? 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.â€?  Background  paper for Afghanistan growth study, World Bank, Washington DC.  Menkhaus, Ken. “Somalia and the Horn of Africa.â€? WDR 2011, Background Case Study.  Suhrke,  Astri.  2006.  “When  More  Is  Less:  Aiding  Statebuilding  in  Afghanistan.â€?  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.â€? 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 conflictâ€?affected provinces and betterâ€?off households.  Thus the largest impact of declining aid on employment and (to a lesser extent) poverty is likely to be  felt  in  conflictâ€?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  aidâ€?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  endâ€? 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 threeâ€?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  postâ€?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  higherâ€?thanâ€?average  growth  immediately  postâ€? conflict.  Although  the  economic  performance  of  postâ€?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 â€?10 â€?20 Agriculture GDP growth â€?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 (securityâ€?related and civilian) were equivalent to nearly  100  percent  of  GDP  in  2010/11.7  Foreign  aid  has  been  on  the  rise  since  2002,  growing  from  US$  404                                                               7  Aid refers here to normal civilian aid as well as securityâ€?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 â€?10 â€?20 â€?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 â€?5 Services Other industry Construction Manufacturing Mining and quarrying Agriculture â€?10   Source: IMF and CSO, World Bank staff calculations.  Note: Excluding opium production.   The fastestâ€?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 postâ€? 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 muchâ€?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 longâ€?term growth.   Moving  away  from  economic  reliance  on  opium  is  thus  a  priority  development  objective.  But  given  opium’s  characteristic as both a highâ€?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  “poppyâ€?freeâ€?  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  “projectionsâ€?  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  onâ€?budget  assistance), or nearly 40 percent of GDP. Donor grants accounted for 50 percent of the core budget in  2010. Offâ€?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 $)   Source: World Bank staff calculations.   Note: Aid includes securityâ€?related aid.  27    Aid  is  delivered  through  three  basic  modalities  (table  2.2).  First,  in  2010  the  government  received  around  US$  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  securityâ€?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).8  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   Securityâ€?related aid  988   1,905   7,028   2,750   5,470   8,594   Onâ€?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  selfâ€?reporting  by  the  largest  donors.  Civilian aid includes expenditure of the U.S.â€?financed Commanders Emergency Response Program. Data is crossâ€?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).9                                                                8  PRTs are civil–military teams that operate in most provinces under the command of a lead nation. They are a wide channel for  the offâ€?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.  9  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 offâ€?budget public consumption 40 30 20 10 Afghanistan 0 4 5 6 7 8 9 10 11 12 GDP per capita (current $)    Source: WDI, World Bank.  Public  investment  (17.8  percent  of  GDP  in  2010/11)  is  much  higher  than  private  investment  (10.6  percent,  including  foreign  direct  investment)  and  both  have  declined  since  the  midâ€?2000s.  Public  investment as a share of GDP is similar to private investment levels in other countries, again mirroring  the country’s aidâ€?dependent economic structure and highlighting the need to reverse these shares.  Spending “onâ€? Afghanistan does not equal spending “inâ€? Afghanistan. Cumulative U.S. spending for the  Afghanistan  mission  is  estimated  to  be  as  high  as  US$    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  inâ€?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.10    Table 2.3 Domestic economic content of aid, 2010/11 (percent)  Aid  On budget  External budget  Securityâ€?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:  ï‚· 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 onâ€?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.  ï‚· Limited  agricultural  growth  (AGRâ€?).  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.   ï‚· 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.   ï‚· Deteriorating security situation (GOVâ€?). 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.                                                                10  Due to lack of data, we were unable to update the local content ratio for offâ€?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    ï‚· Rapid decline in aid (AIDâ€?). Under this scenario, donorâ€?managed (offâ€?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 onâ€?budget aid is unchanged from the BASE scenario.  ï‚· 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 mediumâ€?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+ â€?2.0 â€?4.0 â€?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  onâ€? budget than offâ€?budget spending. While the difference appears marginal, other economyâ€?wide effects— for example, more local employment—could be important.  A  more  rapid  decline  in  aid  (AIDâ€?)  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 sectorâ€? specific, yearâ€?toâ€?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  postâ€?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 â€?20 â€?40 â€?60 â€?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 endâ€?March  35    2011.  Correspondingly,  the  debtâ€?toâ€?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 longâ€?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 highâ€?end segment, unaffordable even to wealthy Afghans, they are expected to  soften. Weaker demand for highâ€?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 existâ€?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  workingâ€?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 selfâ€?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 nonâ€?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  nonâ€?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  shortâ€?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 shortâ€?term in nature.   36    Figure 2.10 Unemployment and underemployment by province    20.0 90.0 18.0 80.0 16.0 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 Kunduz Khost Wardak Kabul Kapisa Farah Faryab Paktika Kunarha Daykundi Nimroz Logar Parwan Ghazni Urozgan Baghlan Laghman Helmand Zabul Paktya Sarâ€?Iâ€?Pul Badakhshan  Takhar Panjsher Nooristan Bamyan Ghor Badghis Nangarhar Balkh Jawzjan Kandahar Samangan Herat Khost Kunduz Kapisa Wardak Kabul Daykundi Kunarha Faryab Paktika Farah Nimroz Logar Ghazni Parwan Helmand Urozgan Baghlan Nooristan Laghman Sarâ€?Iâ€?Pul Zabul Ghor Balkh Bamyan Paktya Panjsher Takhar Badghis Jawzjan Nangarhar Samangan Kandahar Source: NRVA 2007/08.  Note: Yellow bars = nonâ€?conflict provinces, blue bars = provinces with PRTs, blueâ€?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  aidâ€?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 aidâ€?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  aidâ€?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  nonâ€?security ministries received salaries or salary topâ€?ups financed by donors. The International Security  Assistance  Force  registers  60,000–80,000  Afghans  directly  employed  through  militaryâ€?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 secondâ€?round  effects  (jobs  created  indirectly  through  aid  money).  USâ€?CENTCOM  reports  that  U.S.â€?financed  contracts  employed  34,200–78,500  Afghans,  including  contracts  issued  by  USAID  and  military  agencies.11  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 offâ€?budget aid range roughly between 312,000 and 620,000, depending on local  content assumptions.12 Including the number of beneficiaries in the civil service, this suggests that 6.5–                                                              11  USâ€?CENTCOM reports (July 2008–January 2011).   12  We assumed local content of 0.15–0.25 for offâ€?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  securityâ€?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.13  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:   ï‚· 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  (offâ€?budget)  aid  would  affect  11,000–18,000  jobs  (on  a  sixâ€?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.   ï‚· 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 endâ€?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 serviceâ€? (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.  ï‚· 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 conflictâ€?affected provinces  will bear  the  brunt of the  impact as the transition  unfolds, as offâ€?budget  provincial aid allocations  will likely decline faster than those through the core budget.   ï‚· 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).   ï‚· 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 cashâ€?forâ€?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,  highâ€?wage  earners  who  hold  (highâ€? paying) jobs in donor projects. They are more likely to find other jobs in the expanding regular civil  service, though at lower wages.                                                                13    Information derived from USâ€?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  2007–08.14  For  comparative  analysis,  Afghanistan’s  34  provinces  are  classified  into  three  groups  based  on  their  annual  per  capita  spending:  highâ€?spending  (more  than  US$  200);  mediumâ€?spending  (US$  100–US$  200);  and  the  benchmark  lowâ€?spending  group  (less than US$ 100 per capita). To compare their development, we look at four indicators of wellâ€?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.15 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.                                                               14  Matt Waldman (2008): Aid Effectiveness in Afghanistan, Appendix 2 (based on the Ministry of Finance data).     15  We use “number of incidents causing civilian casualtiesâ€? 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 differenceâ€?inâ€?difference approach is used  to  calculate  the  change  in  development  outcomes.  Highâ€?spending  provinces  have  witnessed  a  15 percentage point net decline (compared with lowâ€?spending provinces) in food poverty between 2005  and  2008  (table 2.5).16  But  the  gain  in  per  capita  food  consumption  is  modest  for  highâ€?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, lowâ€?spending provinces  have  outperformed  highâ€?spending  counterparts  in  school  enrollment.  The  weak  impact  of  provincial  spending  on  development  outcomes  is  further  supported  by  the  results  for  the  mediumâ€?spending  provinces,  which  appear  to  have  performed  poorly  on  all  indicators  relative  to  their  lowâ€?spending  counterparts.   Table 2.5 Change in key outcomes relative to lowâ€?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  highâ€?spending  provinces  appears  to  drop  by  another  4 percentage points compared with lowâ€?spending provinces.17 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 mediumâ€?spending provinces shows some improvement, but the results remain worse than  those for their lowâ€?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: highâ€?spending  provinces saw an average increase of Af 70 in their per capita food consumption in 2005–08 relative to  the lowâ€?spending group. This average gain can be further disaggregated by looking at the change in food  consumption  that  accrued  to  the  poor  and  betterâ€?off  provinces  (in  terms  of  their  per  capita  food  consumption ranking; table 2.7). Betterâ€?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 wellâ€?off and the poorer provinces shrinks, however, once the impact of conflict is taken out.                                                                16  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).   17  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 analysisâ€? 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  Betterâ€?off  Increase in per capita food consumption  (25th percentile)  (50th percentile)  (75th 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,18  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  wellâ€?off  provinces.  Some  evidence  also  suggests  that  differential  per  capita  spending  has  contributed  to  inequality of wellâ€?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  “normalâ€? level of aid will present not only risks but also opportunities for Afghanistan to transition to a  more stable, selfâ€?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.                                                                18  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  nonâ€?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.â€? 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 offâ€?budget spending, which is currently less than a third of onâ€?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  longâ€?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.19   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  transitionâ€?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 longâ€?term endeavors. Mitigating the shortâ€?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                                                               19  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.  USâ€?CENTCOM. July 2008–January 2011. “Contractor support of U.S. operations in the USâ€?CENTCOM area  of responsibility, Iraq and Afghanistan.â€? 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  mediumâ€?  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.20 Historically, around twoâ€?thirds of donor  outlays bypass the budget. Of the onâ€?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.                                                               20  Onâ€?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  donorâ€?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 allâ€?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 billion21—about  the  same  as  GDP.  While                                                               21  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,â€? and only US$ 1.9 billion or 12 percent was on budget, through  the  “core  budget.â€?  The  operating  budget  consists  largely  of  the  government  wage  bill  plus  non  wage  O&M costs, funded by domestic revenue, multiâ€?donor trust funds, and donors’ budget support. The core  development  budget  consists  of  donorâ€?financed  projects  and  programs  executed  by  the  government,  plus a small portion funded by budget support and domestic revenue.22 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, CSTCâ€?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  onâ€?budget  contributions  to  the  Afghanistan  Reconstruction  Trust  Fund  (ARTF),  Combined  Security  Transition  Command  (CSTCA),  and  Law  and  Order  Trust  Fund  for  Afghanistan (LOTFA).                                                               22  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 offâ€?budget projects, but the government struggles to align offâ€?budget  programs with national objectives.  49    Wages  and  salaries  account  for  some  threeâ€?quarters of  the  operating  budget  (figure 3.5).    In  2010/11,  the security wage bill alone represented about twoâ€?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.23  After  salaries,  the  secondâ€?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.24  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  onâ€?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.                                                                23  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.  24  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 sustainability25 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 debtâ€?relief efforts under the Heavily Indebted Poor Countries (HIPC) completion point. The joint  IMF–World  Bank  Debt  Sustainability  Analysis  in  2011  found  that  debtâ€?burden  indicators  would                                                               25  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  2010.26  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).   Mediumâ€?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  valueâ€?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).27 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).28   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  secondâ€?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.                                                               26  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 countryâ€?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).  27  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 profitâ€?based income taxes from both mines were based on output and price  estimates that determine gross revenues.  28  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  IMFâ€?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% Stateâ€?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  2021/22.29  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.                                                                29  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 endâ€?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 oneâ€?hour flight in a  military jet stands close to US$ 7,000.30 So a considerable component of the US$ 3.5 billion per year will  be related to fuel.  Nonâ€?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. Nonâ€?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  reâ€?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.                                                                30  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.31   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  onâ€?budget  contributions,  fulfilling  their  Kabul  Conference  commitment  of  shifting  more  of  their  aid  on  budget,  while reducing offâ€?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.                                                                31  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 soonâ€?toâ€?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% â€?5% Operating balance with  security Total core balance without  â€?11% â€?10% security Percentage of GDP â€?16% â€?20% â€?25% â€?30% Total core balance with  security â€?40% â€?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 â€?1,000 â€?2,000 â€?$2,500 â€?3,000 â€?$3,200 2011/12 $ million â€?4,000 â€?5,000 â€?6,000 â€?$5,700 â€?7,000 â€?$7,300 â€?$7,500 â€?8,000 â€?$7,800 â€?$7,700 â€?$7,700 â€?$7,900 â€?$7,900 â€?$8,400 â€?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  lowâ€?income  countries  receive,  on  average,  around  9 percent  of  gross  national  income  in  nonâ€? 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  tradeâ€?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  tradeâ€?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 onâ€?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 hopedâ€?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 nonâ€?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 largeâ€?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 onâ€?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.â€? 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 longâ€?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 partâ€?time employmentâ€? (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.32 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 topâ€?ups via government or donorâ€? 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.33 Most capacity has been built among contracted staff of donorâ€?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  capacityâ€?  (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 buildingâ€? (Michailof 2007: 1). A 2009  OECD report expressed similar views.34   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,35 which announced a new civilian  as well as military surge. Despite this and other international efforts there is little indication that longer  term  capacityâ€?building  objectives  are  much  nearer  to  being  achieved.36  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  longâ€?term  goals  of  building  government  capacity  and  fiscal  sustainability.â€?37  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  recedeâ€? (U.S. Senate Foreign Relations Committee 2011: 3).                                                                32  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.  33  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).   34  “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 governmentâ€? (OECD 2009: 29).  35  Announced in speeches by President Obama on March 27 and December 1, 2009.  36  GAO (2011a) states that the overall impact of U.S. efforts are unknown because of lack of performance data.  37  SIGAR Auditâ€?11â€?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  postâ€?conflict  experience  of  most  lowâ€?  and  middleâ€?income  countries—such  as  Georgia,  Liberia,  Serbia,  Sierra  Leone,  and  Timorâ€?Leste—has  been  similar  to  Afghanistan’s,  especially  in  their  initial  widespread  dependency  on  mostly  ad  hoc,  donorâ€?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 donorâ€?funded programs or NGOs on  much  higher  wages.  An  OECD  report  in  2009,  for  example,  accuses  donors  in  Timorâ€?Leste  of  “undermining the Government’s ability to perform, by engaging in ‘capacity stripping’â€? (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  donorâ€?supported  salaryâ€?supplement  schemes,  as  most skilled civil servants show little interest in working directly for the government.   In Liberia and Sierra Leone, donorâ€?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 topâ€?up schemes to attract key  individuals back from the diaspora.   In  Timorâ€?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  donorâ€?supported  capacityâ€?building  fund it hired staff to support reforms (UNDP 2007). Once domestic revenue had sufficiently recovered,  the government took over the salaryâ€?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 multiâ€?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 longâ€?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 shortâ€?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,  shortâ€?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  lowâ€?income  countries are more constrained. Without longâ€?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  lowâ€?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  ï‚· Donor  capacityâ€?building  interventions  need  to  support  a  common  and  agreed  upon  administrative  reform  plan.  ï‚· Provision of external technical assistance needs to be based on clear assessment of need.  ï‚· Technical  assistance  should  have  builtâ€?in  sunset  arrangements  and  be  tied  to  specific,  defined  objectives  (including “transfer of knowledgeâ€?).  ï‚· Salary  supplementation  schemes  need  to  be  managed  by  the  government  (not  the  donors)  and  made  fully  transparent with published pay scales and allowances.  ï‚· Such  schemes  should  be  part  of  a  broader  strategy  to  address  shortâ€?term  constraints  and  build  longâ€?term  capacity.  ï‚· 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.  ï‚· Salary supplements should be linked to the function rather than the person to allow for more strategic focus.  ï‚· 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’?â€? (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.â€?  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,  meritâ€?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  meritâ€?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 subâ€?national units of government, and ii) to develop core  civil service capacity  both at the centre and at subâ€?national levels to achieve better service delivery and  budget execution outcomes in an increasingly tight fiscal environment.  Subâ€?national administration  Subâ€?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  coâ€?exists  with – and is cut across by â€? 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 agencies38. In the traditional bureaucratic structure, provincial governors were midâ€?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 Subâ€?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  subâ€?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                                                               38  See World Bank (2008), Intergovernment Fiscal Relations and Subâ€?national Expenditures in Afghanistan, August  2008.   67    government structures, locally elected bodies, donorâ€?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)39, which are proposed  at  some  future  date  to  become  Village  Councils40,  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  subâ€?national  administrative  bureaucracy  (especially  the  provincial  departments  of  line  ministries  but  also the Governors’ offices) remains poorly resourced and underâ€?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 decisionâ€?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  subâ€?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 budgetâ€? that it received through the government’s core budget. This is all set to  change. Transition will reduce the flow of these “off budgetâ€? 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  budgetâ€?  resources  should  help  reâ€?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  nonâ€?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 nonâ€?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.41                                                                39  CDCs are villageâ€?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.    40  In the continuing absence of Independent Electoral Commission sanctioned village elections for Village Councils an interâ€? 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.  41  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  nonâ€?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 nonâ€?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  nonâ€?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  nonâ€?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  salaryâ€?support  arrangements.  Several civil service reform programs were carried out in the 2000s.42 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 nonâ€?reform salaries (those that have yet to  go through P&G); interim  additional allowances paid through priority reform and restructuring and its  “superscaleâ€?;  presidential  awards  for  deputy  ministers,  chiefs  of  staff,  and  general  directors;  cadre  allowances for specific jobs such as principals and senior doctors; toppedâ€?up salaries for employees of  revenueâ€?generating  ministries;  senior  staff  employed  through  the  ARTFâ€?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, careerâ€?based civil service to an  open,  meritâ€?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                                                               42  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 meritâ€?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 midâ€?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 nonâ€?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 contractedâ€?in staff paid by donors. Ashraf Ghani, Afghanistan’s former Minister of Finance,  has been a longâ€?standing critic of donors for fueling this “second civil service.â€? 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 servantsâ€? (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 onâ€?budget  government projects, and, on  the  other,  technical  assistance  provided  to  parallel  institutions,  like  PRTs,  and  to  offâ€?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 nonâ€?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  offâ€?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,43  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).44 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: OECDâ€?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  11â€?17â€?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.45   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  conflictâ€?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                                                               43  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).  44  Technical cooperation is defined by OECD/DAC as financing of activities whose primary purpose is to augment the level of  knowledge, skills, technical knowâ€?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 OECDâ€?CRS data.  45  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 capacityâ€?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  shortâ€?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 upperâ€? and middleâ€?  management  positions  in  key  ministries  and  agencies,  offering  “lateralâ€?  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.46 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  reformâ€?minded  leadership  in  a  ministry  with  enough  highâ€?quality  MCPâ€?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.47 Achieving the right  scale of impact will require not only increased, but also longâ€?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.48   In this subsection we look at the costs of those in the second civil service who work either directly for  government (as shortâ€?term contracted civil servants or in largely civil serviceâ€?type roles in line ministries)  or for project implementation units implementing onâ€?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 offâ€?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                                                               46  In his 2009 Atlantic Council article, Ashraf Ghani advocated recruiting up to 3,000 senior civil servants (Ghani 2009).  47  See the conclusions of the Implementation, Completion and Results Report on Afghanistan’s Civil Service Capacity Building  Project (World Bank 2011b).  48  This report, “Donor Funded Remuneration of Individuals Working for Governmentâ€? 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 onâ€?budget payroll of these ministries (figure 4.1).49 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  topâ€?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.50                                                                49  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 lifeâ€?support costs are excluded. 50  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 “harmonizedâ€? 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/ARTFâ€?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  Directorâ€?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  donorâ€?funded  activities,  has  no  tashkeel  civil  servants.  In  MRRD’s  national  programs  like NSP, implementation is almost entirely dependent on contract staff.   Six  ARTF/IDAâ€?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 operatingâ€?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,51 neither donors, policymakers, nor individual Afghans have seen  much  incentive  to  rationalize  their  use.52  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 salaryâ€?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  ringâ€?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                                                               51  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.   52  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 selfâ€?regulation in paying national EFS, and adopt an agreedâ€?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  longâ€?term  capacity,  in  the  future  they  must  be  better  anchored  in  the  government’s  wider  capacityâ€?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  fiveâ€?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  demandâ€?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 highâ€?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  capacityâ€?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  serviceâ€?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 shortâ€?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  serviceâ€?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  ARTFâ€?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 shortâ€?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 demandâ€?driven and  resultsâ€?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  mediumâ€?term  financing  to  frontâ€?line  ministries  to  improve  ministry  efficiency  and  performance;  and  developing  an  incentive  and  governance  environment  that  makes  the  civil  service  attractive  as  longâ€? 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.â€? Paper 7â€?5700. Washington, DC.  GAO (U.S. Government Accountability Office). 2011a. “Afghanistan Governance: Performanceâ€?Data Gaps  Hinder  Overall  Assessment  of  U.S.  Efforts  to  Build  Financial  Management  Capacity.â€?  Paper  11â€? 907. Washington, DC.  ———. 2011b. “Afghanistan’s Donor Dependence.â€? Paper 11â€?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.â€?  Paper  prepared for the World Bank’s Afghanistan Development Forum, April 26.  Ministry  of  Finance.  2010.  “Donorâ€?Funded  Remuneration  of  Individuals  Working  for  Government.â€?  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  Timorâ€?Leste,  Partnership  for  Democratic  Governance  Experts’  Series.  Paris.  ———. 2011. Report on International Engagement in Fragile States, Democratic Republic of Timorâ€?Leste.  Paris.  Rabrenovic, Aleksandar. 2010. Civil Service Professionalization in the Western Balkans. Expert Report for  the  SIGMAâ€?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. Auditâ€?11â€?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. Auditâ€?11â€?17. Washington DC.  Surhke,  A.,  and  Hakimi,  A.  2011.  “Sustainability  and  Transition  in  Afghanistan.  A  Political  Economy  Analysis.â€? 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. Midâ€?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 5â€?306â€?08â€?012â€? P. Washington, DC.  World  Bank.  2004.  Subâ€?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  Lowâ€?Income  Countries under Stress. Washington, DC.  ———.  2008.  Afghanistan—Building  an  Effective  State:  Priorities  for  Public  Administration  Reform.  Report 42166â€?AF. Washington, DC.  ———.    2008.  Intergovernmental  Fiscal  Relations  and  Subâ€?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  Postâ€?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  sectorâ€?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, microâ€?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 lowâ€?income countries in social and physical infrastructure.   82    Development  activity  in  Afghanistan  is  fragmented,  with  funding  spread  across  many  different  and  uncoordinated offâ€? and onâ€?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  &  Marketâ€? 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    Eâ€?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,  costâ€?effectiveness,  and  sustainability.  Service  delivery expansion since 2001 has been largely programâ€?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.53   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:  ï‚· 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).  ï‚· Only  the  district  governor’s  office  has  consistent  support  from  higher  levels  of  government  administration.  ï‚· 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).  ï‚· Electricity supply varies greatly by district but has improved in recent years.  ï‚· They have no procurement capacity.  ï‚· Village  councils  and  shuras,  as  well  as  representative  bodies,  have  proliferated,  duplicating  some  roles.  ï‚· 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                                                               53  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 roadâ€?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  justiceâ€?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  including  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):  ï‚· Governance  and  essential  services  (district  governor’s  office;  justice  institutions;  administrative  staff;  education,  health,  and  police  workers;  and  O&M  costs  of  subâ€?national  administration  and  district, education, and health councils).  ï‚· 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).  ï‚· External budget (other than for roads).  ï‚· 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                       External budgeta  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  rightâ€?most  column  of  table 5.2,  need  to  be  interpreted  with  considerable  caution,  but  suggest:  ï‚· The  rough  estimate  for  onâ€?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.   ï‚· 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 inâ€?depth dataâ€?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 onâ€? and offâ€?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 2011.54  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 governmentâ€?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.   Offâ€?budget education expenditure reached at least US$ 128 million in 1389 (2010/11).55                                                                54  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.  55  Offâ€?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.)  $454 million    Source: Ministry of Finance 2011a.  Education is delivered primarily by civil servants paid from the operating budget (its share is higher than  that of any other sector except security). The bulk of the Ministry of Education’s payroll is made up of  teachers,  most  of  whom  have  undergone  pay  and  grading  reform.  This  reform,  as  well  as  the  introduction of meritâ€?based hiring and promotion, has boosted teacher salaries to an average US$ 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).56 USAID contributed US$ 10 million to the  core  education  budget  and  spent  an  additional  US$  83 million  offâ€?budget  on  communityâ€?based                                                               56  Other donors with onâ€?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 educationâ€?focused NGOs.57  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 $29 million    Source: Ministry of Finance 2011a.  Despite rapid gains in student numbers and the high current expenditure, access to education remains  low in rural areas, particularly for girls. In 2010 enrolled students were estimated to be only 58 percent  of  the  total  schoolâ€?age  population,  and  almost  15 percent  of  those  enrolled  are  permanently  absent.  Variations across provinces in education outcomes are wide, with primary net enrollment rates ranging  from  30 percent  in  some  areas  to  almost  100 percent  in  others.  The  variation  in  girls’  primary  enrollment is especially wide, reportedly ranging from less than 10 percent in Helmand and Uruzgan to  almost  90 percent  in  Nimroz  province.58  The  current  school  population  is  also  heavily  concentrated  in  grades 1–4, with high dropout rates in higher grades.  Widespread  concerns  exist  over  education  quality,  owing  to  the  poor  qualifications  of  some  teachers,  lack  of  a  standardized  curriculum,  lack  of  evaluation  standards,  and  insufficiency  of  basic  school  supplies.  Statistics  collected  in  2009–10  by  a  national  teacher  registration  system  indicate  that  only  27 percent of the 162,000 registered general education teachers are educated at a grade 14 level59, the                                                               57  Other large donors supporting offâ€?budget education programs are Germany (US$ 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.   58  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.  59  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.60 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 highâ€?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  offâ€?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.                                                               60  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 lowâ€?income country  averages (WHO 2011).61  But public spending is low relative to the population’s needs, and  the  private  sector  serves  a  disproportionate  role  in  health  care  provision,  with  outâ€?ofâ€?pocket  payments  to  health  care providers accounting for threeâ€?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  villageâ€?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 donorâ€?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.62  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  grantsâ€?  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  semiâ€?autonomous  Agriculture  Development  Fund,                                                               61  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.  62  Of block grants disbursed by NSP from 2003 through midâ€?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 nonâ€?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.  Onâ€?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 offâ€?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  offâ€?budget  donor,  the  European  Union,  contributed  US$  66 million  to  a  range  of  programs  on  food  security,  agricultural  production, and riverâ€?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 $37 million    Source: Ministry of Finance 2011a.    NSP,  which  provides  block  grants  to  villages  for  smallâ€?scale  projects  prioritized  by  elected  Community  Development  Councils,  has  been  one  of  the  most  successful  national  programs.    It  has  reached  some  26,000 villages, and more than 45,000 subâ€?projects have been completed, in every province and district  of  Afghanistan.  When  NSP  was  designed,  the  initial  block  grants  were  deliberately  set  at  a  high  perâ€? capita  level,  reflecting  the  long  neglect  of  rural  development,  urgent  local  reconstruction  priorities,  higher costs of construction materials, and in particular the need for government to make a significant  impact in rural areas to enhance its legitimacy.  According to NSP data the average value of block grants  is US$ 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.63                                                                    63  Given lack of data on village populations, and perhaps some confusion about the definition of a “householdâ€? 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.64  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 Mazarâ€? 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.65 Although  some have been carried out offâ€?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  communityâ€?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  offâ€?budget  funding  for  road  construction,  rehabilitation,  resurfacing,  and  maintenance  through  USAID’s Infrastructure Rehabilitation Program.                                                               64  Information sources for this subsection include USAID (2010 and 2011a,b) and Altai Consulting (2011a).   65  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  performanceâ€?based  management  model  that  assigns  sections  of  road  to  local  firms,  which  assume  responsibility  for  routine  and  emergency  maintenance  to  ensure  a  preâ€?established  level  of  service.  In  2009,  more  than  US$  200 million  was  collected  in  transportâ€?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  performanceâ€?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.66 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 Mazarâ€?eâ€?Sharif having access to 24â€?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.67  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,68 but insecurity, drought, logistical challenges, and crossâ€? border riparian issues have hindered efforts to rehabilitate or build hydroelectric plants. Hydroelectric                                                               66  Information sources for this subsection include the Ministry of Energy and Water (2007), Da Afghanistan Breshna Sherkat    (2011), Interâ€?Ministerial Commission for Energy (2010 and 2011), and Altai Consulting (2011a).  67  Similar investments in western Afghanistan have improved transmission and distribution systems, linking urban centers to  sources in the Islamic Republic of Iran and Turkmenistan.  68  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.69  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.70  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 longâ€?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 largeâ€?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 $51 million   Source: Ministry of Finance 2011a.                                                               69  The cost of imported electricity is US$ 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  offâ€?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  Interâ€?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  kilowattâ€?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.71 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  sectorâ€?specific  difficulties  due  to  diversity  in  implementation  modalities and levels and sources of funding.                                                               71  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  transitionâ€?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  offâ€?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 NGOâ€?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  lifeâ€?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  forâ€?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  microâ€?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  villageâ€?level  infrastructure  built  under NSP and other communityâ€?based programs,72 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).                                                               72  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 underâ€?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  transportâ€?related  revenue  collection  and  contract  maintenance works using a performanceâ€?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 lowâ€?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 transmissionâ€?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 costâ€?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 highâ€?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 largeâ€?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 sectorâ€?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  sectorâ€?specific  strategies  will  be  needed  to  manage  and  mitigate  sectorâ€?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 shortâ€?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 ruleâ€?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  Beastonâ€?Blaakman,  A.,  and  M.  Cros.  2011.  A  Costâ€?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.  Interâ€?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.â€? Wall Street Journal, July 13.  USAID  (U.S.  Agency  for  International  Development).  2010.  “Officials  to  Establish  Road  Authority.â€?  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.â€? 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.â€? 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 mediumâ€? to longer term perspective. These include:  ï‚· Economic impacts: growth prospects, structural change, employment, poverty, and future economic  drivers (chapter 2).  ï‚· Fiscal challenges: expenditure liabilities, large fiscal gaps and how to close them, revenue prospects,  and absorption of more aid onâ€?budget (chapter 3).  ï‚· Government  capacity  and  how  to  boost,  consolidate,  and  sustain  it  during  transition:  rationalizing  the  “second  civil  serviceâ€?  and  transforming  it  into  sustainable  core  government  capacity,  and  continuing public administration reforms (chapter 4).  ï‚· 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 runâ€?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  governmentâ€?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  longâ€?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  onâ€?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  “onâ€?  Afghanistan  is  not  spent  “inâ€?  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  (onâ€?budget  aid  has  a  much  higher  impact  on  the  Afghan  economy  than  offâ€?budget  assistance  does).   Underemployment  will  rise  because  the  activities  affected  by  declining  financial  inflows  (services,  construction)  are  relatively  laborâ€?intensive.  Roughly  6–10  percent  of  the  working  population  has  benefited  from  aidâ€?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,  conflictâ€?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  conflictâ€?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  shortâ€? run  stabilization  activities.  Aid  disproportionately  devoted  to  conflictâ€?affected  provinces  has  only  modestly impacted poverty. Moreover, households in the conflictâ€?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 hopedâ€?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  nonâ€?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  lowâ€?income  countries  receive  around  9 percent  of  gross  national  income  in  nonâ€?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  donorâ€? 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  onâ€?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 serviceâ€? 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 foreignâ€?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  economyâ€?wide  simulation  model  created  to  analyze  development  strategies  and  their  impact on economic growth.73 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                                                                73  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  activityâ€?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.74  An  assumed  share  of  the  mining  capital  adds  to  public  capital,  reflecting  the  benefit  of  miningâ€?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.                                                               74  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 nonâ€?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.75   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                                                               75  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  mediumâ€?term projection but is otherwise endogenous for the remaining years.76 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  2011/12–2025/26.77  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  (onâ€?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.                                                               76  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 sectorâ€?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.  77  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  onâ€?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.78  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.                                                                78  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,79  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.80 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), (onâ€?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.                                                                79  This means that for every 100 Afghanis worth of public investments total value added will increase by 14.2 the following year.  80  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  â€?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, AGRâ€?  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  AGRâ€?  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  slowâ€?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 onâ€?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 (onâ€? budget) from donors (external budget).81 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 onâ€? 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 onâ€?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  Governmentâ€?managed     Nonsecurity  69  Security  96  Donorâ€?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                                                               81  The agreed objective at the Kabul conference in 2011 was to shift 50 percent of donor funding onâ€?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 onâ€?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, AIDâ€?  Under  this scenario, donorâ€?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  Onâ€?budget  aid  and  external  aid  developments  in  the  AIDâ€?  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.82  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                                                               82  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  AIDâ€?  and  the  BASE  by  2025/26,  but  it  was  more  drastic  in  the  beginning  for  AIDâ€?  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  donorâ€?run  consumption  and  investment,  which  is  highly  importâ€?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  wellâ€?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, GOVâ€?  GOVâ€?  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 GOVâ€? case, annual average GDP growth is even negative—that is, GDP decreases during 2010/11– 2025/26. GDP growth is â€?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.83   Domestic  demand  (absorption)  declines  even  more  (â€?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.                                                               83  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 GOVâ€? 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 GOVâ€? 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  Onâ€?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 Diazâ€?Bonilla, and Hans Lofgren, 2008. “Aid, Service delivery and the  Millennium Development Goals in an Economy wide Framework,â€? pp. 283â€?315 in Francois Bourguinon,  Maurizio Bussolo, and Luiz A. Pereira da Silva, eds. The Impact of Macroeconomic Policies on Poverty  and Income Distribution: Macroâ€?Micro Evaluation Techniques and Tools.  Washington, D.C., World Bank.  Lofgren, Hans, Martin Cicowiez, and Carolina Diazâ€?Bonilla. 2012. “MAMS – A Computable General  Equilibrium Model for Developing Country Strategy Analysis.â€? 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 profitâ€?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.84 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 midâ€?range operating costs globally. The  estimated cost of pellet production at Hajigak is US$ 75 per tonne, which is based on midâ€?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.85 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.)                                                               84  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.  85  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.86 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.87  A  nonsignificant  correlation  between  foodâ€?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 aidâ€?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.  Valueâ€?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.                                                                86  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).  87  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 soonâ€?toâ€?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 Assessmentâ€?.  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  endâ€? 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,  topâ€?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 highâ€?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.  Meritâ€?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.  Meritâ€?based  recruitment  Salary topâ€?ups  Ministryâ€?specific  topâ€?ups  to  Ministry  and  Revenueâ€?generating  ministries  can  use  up  to  civil servants. Performanceâ€? or  donor specific  5 percent of their revenue to provide topâ€?ups  incentiveâ€?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  Â