The Afghan War: Key Developments and Metrics

The United States is now reexamining its strategy for Afghanistan, and whether it needs to shift from a “deadlines-based” strategy tied to steady reductions in its forces and effort to a “conditions-based” strategy designed to meet Afghanistan’s actual needs. This raises critical issues regarding the success or failure of past U.S. efforts to create effective Afghan forces, and the conditions in Afghanistan that will help determine the probable success of any new U.S. strategy.

A new report by the CSIS Burke Chair addresses these issues in depth by comparing the metrics on the Afghan War that are available to illuminate the current conditions in Afghanistan. This report is entitled The Afghan War: Key Developments and Metrics and it is available on the CSIS web site at

The report provides a survey of the metrics on several key aspects of the Afghan war. These include the cost of the war, the trends in the fighting, the nature of the U.S. train and assist mission, the progress and problems shaping Afghan security forces, the lack of resources for civil-military operations, and the problems in Afghan governance, corruption, and popular support.

The survey also covers the levels of stress in Afghan economics and business development, in social dynamics and poverty, and the impact of Afghan demographics, the resulting youth bulge, and growing urbanization on the war. The final section flags the growing importance of narcotics to the Afghan economy and each side in the war.

Such a report on the metrics of the war can only tell part of the story. Metrics are limited by the quality of the data, the ability to find a credible source of the most useful information, the need for security or to serve political goals, and by the ability to quantify or map the data involved. They need narratives to put them in context, and there are many areas—such as the strategy and politics of a conflict—where metrics can only provide limited insights.

This survey cannot avoid such limitations, but metrics do have the advantage that they can often summarize complex patterns in the conflict in ways that narrative cannot, and can flag areas where the trends raise major issues for policy planning, strategy, and analysis. They also provide a way of comparing very different trends and aspects of a conflict and flagging interactions that might otherwise be far less apparent.

The key sections of the report include metrics that provide the following insights into the challenges that the United States, its allies, and Afghanistan face in formulating and implementing a new strategy:

The Size and the Cost of the Current U.S. Effort

One of the stranger aspects of America’s current wars is that no administration or element of Congress has ever provided an official estimate of the cost of each war. The closest thing to such an estimate is the work of Amy Belasco of the Congressional Research Service (CRS), and this work relies on partial reporting that may not fully include important medical and retirement costs, and leaves serious issues about what is and is not included in the DoD and State Department OCO and other budget data it draws upon.

The tables and graphs in the report show that this estimate put the cost of the Afghan War at $743.7 billion from FY2001-2003, with $700.7 billion going to DoD (current military costs less medical and retirement) (94.2%), and $36.2 billion going to State (4.9%). If one includes the OCO costs for FY2016 and FY2017 reflected in the revised DoD estimate for March 2017 ($129 billion), and the State Department request for FY2016 and FY2017 as of February 2017 ($35.6 billion), the total current cost of the war through FY2017 would be $901.5 billion.

Unfortunately, the OCO budget for Afghanistan includes spending on Baseline activities as a way of avoiding the budget caps under the Budget Control Act and “in theater support” costs that do not always seem tied to the Afghan War. It has become a catch-all category for manipulating the DoD budget to avoid budget caps, which explains part of the reason a low level of U.S. combat activity in Afghanistan is projected to cost $8 billion in FY2017 and the fighting in Iraq and Syria is costed at only $9 billion.

At present, both the Administration and Congress have failed to provide a credible database for costing the Afghan War, and providing a reliable basis for estimating the trends in these costs.

Conflicting Patterns of Threat Analysis: Estimates of the Threat, Levels of Violence, and Comparative Control of Districts

The U.S. and ISAF were slow to publicly recognize the growth of the Taliban and other threats, and never publicly acknowledged the key role Pakistan and its ISI played in giving threat forces a sanctuary and support. By 2006, however, the metrics showing the rising number of attacks made it clear that the Taliban had returned in force and that ISAF and the Afghan government (GIROA) faced a major threat.

The resulting public analysis did not focus on areas of influence and the political dimensions of what had become a major insurgency, and never provided clear estimates of the size of threat forces, their structure, their effectiveness, and their financing, and often failed to explain the seasonal patterns in the fighting.

It did, however, provide relatively honest and detailed metrics on the patterns in the fighting through CY2009-CY2012. After that point, metrics became more and more limited, and the categories were increasingly tailored to “spin” the outcome in favor of ISAF and the ANSF—gradually approaching the propaganda levels common in the “Vietnam Follies.”

By the time the surge in U.S. forces had ended, a sharp set of difference had emerged between US/ISAF and UN, IMF, and other reporting on the level of threat influence and risk in maps of Afghanistan, and the U.S. and ISAF began to deliberately choose metrics that favored the ANSF and ISAF.

These included limiting reporting to Enemy Initiated Attacks (EIA), and sometimes only effective Enemy Initiated Attacks, which sharply understated the overall level of violence, and ignored the fact that the threat had little reason to directly engage ISAF and ISAF supported forces once the United States had announced that its combat forces were leaving at the end of 2014.

Initially relatively honest metrics on the impact of the surge in U.S. forces—showing a range of patterns rather than just EIAs—were dropped when it became apparent that it had no lasting impact and the threat continued to grow.

The United States and ISAF then ceased to publicly report most other metrics showing the course of the fighting, and new metrics on the transition to Afghan forces gave exaggerated credit to Afghan forces taking the “lead” in what were often minor operations and in ones that had substantial U.S. and ISAF support on the ground and in the air.

By the end of 2016, some reporting bordered on the absurd. It treated all EIAs as the same, and the steady growth of threats to targets like two provincial capitals was disguised by reporting every attack as if it had the same impact.

The graphics showing the level of threat by district were dropped from DoD and ISAF reporting and replaced by changing tables that focused on control of districts in ways where "control" and "contested" were never defined, and where the growth of threatening political influence and control of the countryside seems to have been sharply understated.

Similarly, estimates by the UN and NGOs and by the Taliban seem to reflect a much higher level of threat, as do most of the metrics in the following section of this report. Put bluntly, DoD reporting seems to have limited credibility at best, and Resolute Support (the replacement for ISAF) has not provided any meaningful metrics on the threat.

Conflicting Patterns of Threat Analysis: Afghan Perceptions, Civilian Casualties, and Terrorist Incidents

Three further sets of metrics raise serious issues regarding current official DoD reporting on the current size of the threat: Popular opinion, civilian casualties, and terrorist incidents:

  • A 2016 poll by the Asia Foundation is reassuring in that it shows a steady decline in popular support for the Taliban and other threat forces—although the survey occurred largely in more secure areas under government control. It also showed that those surveyed felt threat forces were fighting for power and not for Islam or the Afghan people.
  • However, the same survey showed mixed regional perceptions of the increase or decrease in the threat during 2015-2016 by region. It also, however, showed a steady increase in fear for personal safety, and high overall levels of fear by region. At the same time, it showed far more fear of encountering the Taliban and Daesh than of encountering the ANA and ANP—although fear of encountering Western forces was surprisingly high.
  • UN civilian casualty estimates roughly doubled between 2009-2016, and ground engagements have become the primary cause of such casualties. Threat forces dominate civilian casualties (61%), but pro-government forces are a rising cause (24%).
  • Civilian casualties have risen sharply in many regions since 2009 and have risen to pre-surge levels in the southern region.
  • The number of terrorist incidents reported in the START database used by the U.S. State Department increased by more than four times between 2011 and the end of 2015.

Shaping and Resourcing the Train and Assist Mission

The metrics on the train and assist mission are complex, but reveal several critical things about the nature of the train and assist mission in Afghanistan:

  • The war in Iraq put constant pressure on U.S. troop levels from 2003-2009, and limited the troops available to U.S. forces in Afghanistan along with the ability to provide any—much less qualified —trainers.
  • The surge in Afghanistan came and went before the effort to create effective ANA and ANP forces was properly funded and supported with trainers.
  • The initial deadlines for withdrawal called for by President Obama did not reflect the military realities on the ground or the trends in the threat.
  • The U.S. did begin to appropriate massive rises in security aid to the ANSF beginning in FY2011. These funds dropped sharply in FY2013, however, before U.S. combat forces left.
  • The key source of Afghan forces aid—the Afghan Security Forces Fund—was highly erratic and dropped sharply in 2012. Funding was not initially sized to deal with a serious increase in the fighting after the transition and departure of U.S. combat forces.
  • The U.S. shifted to a strategy based on fixed deadlines for withdrawal in May 2014, and one whose deadlines took no real account of the realities of the threat, the patterns of conflict, and the real-world level of development in Afghan forces.
  • The military train and assist personnel that remained after 2014, and remaining combat elements supporting the Afghan counter-terrorism force, were too small to even fully cover Afghan forces at the Corps level and were not sized to provide support to ANA combat units (Kandaks), the ANP forces in the field, and the ALP—in part because of fear of attacks by rogue elements in the ANSF, in part to limit costs, and in part to avoid casualties.
  • These numbers were misleading in two respects. They did not count some temporary duty (TDY) military and, from the spring of 2015 onwards, there were an average of three times as many U.S. and Afghan contractors funded by DoD as there were total U.S. military personnel.
  • The train and assist effort was complicated by creating German zones in the north, Italian Zones in the West, Turkish zones in the capital and joint zones with Poland in the east. There was serious attrition in every allied T&A effort during 2016. Personnel dropped from 15,055 in May 2016 to 12,611 in November 2016.
  • The revised plans that eased the deadlines for cutting the T&A mission still left far too few personnel. U.S. military personnel for Afghanistan totaled 10,012 in FY2015 and 9,737 in FY2016. Plans to cut them to 6,217 in FY2017 were canceled, but only 8,674 remained.
  • These numbers do, however, ignore 62,485 in-theater support personnel that support the Afghan, Iraq, and Syria war and all contractors. Contractor personnel remained high. There were 26,022 Total (9,474 U.S.) in Q1 FY2017, and the annual cost in FY2017 was $107.6 million.
  • Active U.S. air support dropped roughly in half in 2015 although the Afghan Air Force had very limited mission capability. Close air support sorties dropped from 12,978 in 2014 to 5,774 in 2015 and 5,162 in 2016. Close air support sorties that dropped a weapon were cut from 1,136 in 2014 to 411 in 2015 and 615 in 2016.

Progress and Problems in Afghan Forces

  • The rising threat led to setting higher and higher Afghan force goals in 2004-2016: 40,000 to 80,000 in February 2008, 134,000 in September 2008, and 171,600 in January 2010.
  • The total goal for the ANA and ANP nearly tripled between July 2007 and April 2011.
  • This rapid expansion made each ANA Corps something of a hollow force through at least April 2012, and no reliable count exists of “ghost” soldiers.
  • Critical shortfalls existed in trainers, and especially skilled trainers, through at least March 2012. No open source reports seem to exist on the actual level and expertise of train and assist personnel relative to requirements after that time.
  • The metrics on training and force development provided through 2012-2013 did give some picture of the level of training and resource given combat units had, but at no point was the actual nature of the level of training—and the real-world level of counterinsurgency training—made public.
  • From the start to the present, the training cycles for most forces were far too short, attrition high, and corruption has presented serious challenges at every level except for ANA Special Forces, the ANCOP forces in the ANP, and elite counterterrorism forces.
  • Reporting on transition exaggerated the number of times the ANSF took the lead in meaningful operations, and reporting was provided on combat effectiveness versus factors like manning, equipment, and force generation training.
  • New problems emerged in 2013-3014 when insider attacks by Afghan forces on trainers and advisors led to serious cuts in forward training and advisory efforts.
  • Afghan force goals stabilized at 195,000 ANA, 157,000 ANP, and 30,000 ALP during FY2015-FY2017. No public explanation was ever given of the rationale for sizing these totals or what kinds of capabilities and forces they were supposed to include.
  • Afghan popular perceptions of the ANA and ANP, however, dropped sharply during this period. Perceptions of police corruption exceeded 50% in half the provinces in Afghanistan in 2016, and perceptions of who actually provided security were mixed.
  • The cost of most U.S. aid to Afghan forces stabilized at around $4.1 billion a year during FY2015-FY2017.The Afghan share of ANSF costs was under 10% in FY2017.
  • The graphics in this and the following sections of this report show that adequate funding and numbers of trainers were not provided until 2011-2013, and that U.S. and allied combat forces were withdrawn for political reasons at least several years before Afghan forces—that had been rushed into existence—had the experience they needed.
  • New problems emerged in 2013-2014 when insider attacks by Afghan forces on trainers and advisors led to serious cuts in forward training and advisory efforts.

Stress in Governance, Corruption, and Popular Support

  • The World Bank provides governance indicators that show that the Afghan government has made little real progress in providing effective and honest governance since 2001. The data on corruption, stability, and rule of law explain why the following data on popular perceptions of the government by the Afghan people are often so negative.
  • Anti-corruption efforts have had no meaningful impact on perceptions of the level of corruption, and the Afghan forces lack an effective and honest partner at every level from national to local.
  • The number of Afghans feeling the country is moving in the right direction dropped from 58% in 2013 to 29% in 2016. The feeling that the country is moving in the wrong direction is shared by all major ethnic groups.
  • Corruption and unemployment are seen as critical problems by roughly the same number of Afghans that see insecurity as their key problem.
  • Satisfaction with government performance has dropped sharply at every level of governance since 2014. So has popular confidence in the various levels of governance.
  • The economic trend data that follow indicate these problems are likely to grow steadily worse and the final, data in this section warn that over 1 million Afghans will be forced back into Afghanistan and out of Pakistan and Iran in 2016-2017.

Funding Military vs. Civil-Military Operations

The graphics in this section help illustrate several key issues:

  • The U.S. and ISAF never funded a meaningful nation building effort in Afghanistan or Iraq, and never provided meaningful training for the elements of the ANSF for civil-military operations.
  • Neither UNAMA nor the U.S. ever developed effective integrated plans for civil-military operations in Afghanistan.
  • As the ESF graph shows, resourcing of the aid effort in Afghanistan went through massive swings, was project-oriented, and was often tied to plans that understated the scale of the fighting or assumed it would soon be over.
  • Governance aid had limited impact, often trying to change or modernize systems that were corrupt, and ignoring popular needs.
  • The Afghan government repeatedly pledged reform plans it did not fully implement.
  • UNAMA, the body that should have coordinated aid, became a divided political nightmare that never produced meaningful plans or coordination efforts.
  • ANA, ANP, and ALP units often lack support from effective civil government at the District and sometimes the Provincial level.
  • In general, Afghan public support of the government has declined steadily since U.S. and ISAF forces have left and the fighting has intensified.

Levels of Stress in Economics and Business

  • Business sentiment remains suppressed.
  • Agriculture sector’s performance was also mixed in 2016; cereals production recorded a decline of nearly 5 percent while fruit production has been higher.
  • Real gross domestic product (GDP) growth is projected to have only marginally increased from 0.8 percent in 2015 to 1.2 percent in 2016. With a population growth of nearly 3 percent, such a level of economic growth implies a decline in per capita income.
  • Inflation increased from -1.5 percent in 2015 to 4.4 percent in 2016, driven by lagged effects of currency depreciation and a recovery in global food prices.
  • Revenue collection has significantly improved in the past two years after the abrupt decline in revenues in 2014. But, revenues were only $1.7 billion and expenditures were $6.639 billion (2015 est.) The budget deficit was 26.8% of GDP (2015 est.).
  • Domestic revenues increased by nearly 15 percent in 2016, which exceeded the revenue target by around 5 percent.
  • Both tax and non-tax revenues increased, while customs duties remained flat given weak imports. In proportion to GDP, however, revenue collection still remains relatively low at 10.7 percent.
  • With an increase in exports and slower growth for imports (due to weaker domestic demand), the trade deficit is estimated to have improved from -36.7 percent of GDP in 2015 to -35.0 percent in 2016.
  • The large trade deficit continues to be financed by foreign aid, with the current account balance expected in a small surplus estimated at around 4 percent of GDP in 2016. (Exports were $658 million (2014 est.) $2.679 billion (2013 est.) not including illicit exports or reexports. Imports were $7.004 billion (2014 est.) $12.19 billion (2013 est.).)

Levels of Stress in Social Dynamics and Poverty

The World Bank provides a far more realistic picture of the pressures on the civilian population than many competing sources. Some other sources seem to deliberately avoid addressing the full economic impact of the war, and the massive cuts in U.S. and allied military spending in country since 2013, as well as cuts in aid.

The World Bank warns that poverty has been increasing since 2008, and rose sharply between 2013 and 2014. There evidently has been limited data on many aspects of these issues since 2014-2015, but the World Bank data still warn that the post-2014 period has been one with:

  • A critical increase in unemployment.
  • Growing problems in the education sector.
  • Similar concerns about health—flagging the fact that some estimates of post-2011 improvements in health were probably grossly exaggerated.

Demographics, Urbanization, and the “Youth Bulge”

The economic strains on “hearts and minds” are further driven by major population pressures.

  • Afghanistan’s deep ethnic and sectarian divisions present serious problems in unifying the country and creating divisions in the ANSF.
  • It has extraordinary population growth and a very high dependency ratio (87%) on those working or with farms and businesses.
  • The youth bulge is acute: Median age is only 18.6 years, 41% is 0-14 years and 22% more is 15-24 years.
  • This youth bulge and dependency ratio are the highest in the region.
  • 71% of youth saw unemployment as their greatest problem as early as 2014 vs. 10-12% who focus on insecurity.
  • Population pressure and security problems have created a massive increase in urbanization, slums, and urban unemployment—potentially destabilizing population centers.
  • 1 in 4 Afghans is either unemployed or under employed and 28% or more of male youth is directly unemployed.

Accepting a Narco-Economy?

Virtually all reporting on the Afghan economy is fundamentally dishonest in that it does not estimate the economic impact of the “illicit” sectors of the economy: corruption and narcotics.

Massive increase in drug growing and output since 2014, and drug eradication programs have clearly been a wasteful failure.

Most reporting of the economic impact of narcotics is meaningless because it only considers farm gate economics, and not the overall impact of narco-processing and trafficking.

Estimates of economic impact are uncertain, but:

  • Some working estimates indicate narcotics may be worth $1.6-$3 billion a year at the national level in 2016.
  • UN estimates that even farm gate prices put Opium at $898 million in 2016.
  • CIA figures indicate that total legal exports were only $658 million 2014, even at farm gate prices.
  • CIA estimates total GDP at $64.08B in PPP terms and at only $18.4 billion in exchange rate terms.

Job estimates for drugs go to 400,000—more than the total ANSF authorized strength.

UN maps show drug growth and sales are key sources of revenue to the Taliban and powerbrokers.

Photo credit: JAVED TANVEER/AFP/Getty Images
Anthony H. Cordesman

Anthony H. Cordesman

Former Emeritus Chair in Strategy