Measuring Success: Are We Winning? Over 10 Years in Afghanistan

This was a report I wrote for a two-year fellowship I had with the American Security Project, a small think tank in Washington, DC. You can find the formatted PDF here, and I’m posting the open text below so it’s easier to read.


For the last two years, the war in Afghanistan has taken a very different turn. While it used to be defined by grinding summers of battles between insurgents and Coalition troops and long winters of relative quiet, the Afghan insurgents have spent the last two years on a sophisticated campaign of assassinations, attacks against high-profile targets in Kabul (which many considered secured and unassailable), and a sophisticated influence campaign.

Starting in January of 2011, they launched a brazen suicide assault on an upscale supermarket in a secured area of Kabul, targeting the country director of a private security firm.1 In February, suicide attackers exploded bombs at a hotel in downtown Kabul, killing two.2 In April, insurgents attacked Afghan Army and International Security Assistance Force bases in the city, though they killed no one other than themselves.3

In June, gunmen launched a sustained, complex attack on the heavily guarded Hotel Intercontinental in Kabul, killing eighteen people.4 Over the summer, they launched a string of deadly attacks, killing dozens.5 In August, insurgents launched yet another complex, sustained attack on the British Council in a wealthy, secured neighborhood of Kabul, killing eight.6 In September, a small team of insurgents managed to fire machine guns and launch RPGs at the U.S. embassy for over 20 hours before being neutralized by Coalition forces.7

In October, an insurgent detonated a car bomb next to an armored bus in Kabul, killing at least five troops and eight U.S. civilians.8 In December, a wave of Taliban bombing and shooting attacks targeted at transitioning Afghan forces killed 13.9 In January 2012, another wave of Taliban attacks targeted Afghan government officials and killed 20.10 In April, insurgents launched a sustained attack on embassies, military bases, and the Afghan parliament.11 And in May of this year, insurgents attacked Kabul during President Obama’s visit, killing 712 – they also killed five education officials in Paktika province near the border with Pakistan.13

Given the persistent underreporting of violence in Afghanistan, there is almost certainly even more violence than this short list captures.14

This marked shift would seem to indicate the war is being lost. But because the U.S. fails to monitor crucial aspects of the war, there is no reliable way to be certain.

10 years after U.S. forces invaded Afghanistan, we still lack the means to tell whether the war is being won or not.

None of those attacks, taken in isolation, killed very many people—the deadliest was the Intercontinental Hotel attack, which included seven dead insurgents — and they weren’t meant to. Even September’s attack on the U.S. embassy wasn’t very complex, and the insurgents did not demonstrate any particular tactical genius.15

The insurgents, however, do not need to win large tactical victories in order to win the war: they are not fighting a symmetric war. Rather, they seem intent on disrupting the Afghan government’s ability to govern - so that rule of the Islamic Emirate of Afghanistan becomes preferable to Afghans.

In 2008, an attack on the U.S. embassy in Kabul was widely considered unthinkable. The downtown neighborhoods were being surrounded by what many called a “Ring of Steel”16 — dozens of guards, roadblocks, checkpoints, and barricades — so that business could take place in a Green Zone where violence was kept at a minimum. Today, the insurgency can penetrate the Ring of Steel again and again to launch attacks.

The recent attacks served to shake confidence in ISAF and the Afghan government.

In contrast, ISAF seems to be waging a different war altogether.

ISAF reporting on its activities seems more concerned with establishing a running tally of operations— relying on insurgent body counts to underscore progress in the war,17 publishing data only on deadly attacks,18 tangible measurements of progress like counting the number of community shuras it hosts,19 and counting how many soldiers finish basic training. 20

While ISAF data captures many physical measurements of the war, it is less successful in measuring the social and political effects of the war, such as which side Afghans think will ultimately win.

In essence, the Taliban and its allies must shake confidence in the Afghan government and diminish the Western appetite for staying in order to win. The International Security and Assistance Force and its Afghan allies are waging a far different war – based on sweeping operations, infrastructure creation, and security force training.

With both parties to the conflict fighting different wars with different types of outcomes, is it even possible to gauge if the ISAF and Afghan government coalition is winning? The answer is not simple.

When gauging ISAF’s success in Afghanistan one should first define success — something we show is not easy to do. Once success is defined, one can create the metrics by which one would measure progress toward or away from that defined success.

We settled on the most generous interpretation of President Obama’s publicly stated aims for the war: deny al Qaeda safe haven, prevent the Taliban from overthrowing the government, and build up the Afghan security forces and the government so they can take responsibility for their country’s future.

How one goes about reaching this goal is also no simple matter: two of the three objectives President Obama stated are defined by absence rather than by achievement, and the third objective is not really definable in concrete terms.

While it may be difficult to say when, exactly, we can know that our goals have been met, we can develop a series of metrics that should indicate whether we are progressing toward those goals or not.

The following paper describes those metrics, and tries to determine whether we are measuring the most relevant data to gauge success in Afghanistan.

Lastly, we ask the question: what are we actually measuring? And does the data we have allow us to accurately gauge the success of our efforts?

What Is Success?

The biggest problem in asking whether we are winning in Afghanistan is defining what winning looks like. What are the aims of the war?

Since taking office, President Obama and his senior officials have offered a variety of definitions for success. In his speech to West Point announcing the military surge, President Obama said:

Our overarching goal remains the same: to disrupt, dismantle, and defeat al Qaeda in Afghanistan and Pakistan, and to prevent its capacity to threaten America and our allies in the future.

To meet that goal, we will pursue the following objectives within Afghanistan. We must deny al Qaeda a safe-haven. We must reverse the Taliban’s momentum and deny it the ability to overthrow the government. And we must strengthen the capacity of Afghanistan’s Security Forces and government, so that they can take lead responsibility for Afghanistan’s future.21

In his State of the Union in 2011, he went on to say: “Our purpose is clear -- by preventing the Taliban from reestablishing a stranglehold over the Afghan people, we will deny al Qaeda the safe-haven that served as a launching pad for 9 /11.” 22

These goals do not easily lead to defined success metrics. Rather they are linked to proving negatives: are we denying them a safe haven, and are we degrading their abilities? These goals create a strategy defined by absence, not progress, and thus make that strategy impossible to achieve.

At any point, a single variance from absence might indicate failure, so the status quo must be maintained forever to ensure there is never variance.

Winning and success requires a goal. And the only way to determine whether that goal has been met is to establish metrics by which you measure your progress toward achieving that goal. There is, however, very little agreement among senior officials in the Obama Administration about what that goal really is:

Ryan Crocker, the Ambassador to Kabul, has said, “Building an Afghanistan that is a shining city on a hill ... is not going to happen” . Instead, the goal should be “sustainable stability.”23 He said our goal is to bring, "‘good enough governance’ in Afghanistan to ensure the country doesn’t ‘degenerate into a safe haven for al-Qaeda.’"24

Former Secretary of Defense Robert Gates tried to define success in Afghanistan without a conventional military victory. “We have not had a declared victory in a war — with the possible exception of the first Gulf War — since World War II,” he said in June last year. “It is the phenomenon of modern conflict.” 25 Secretary Gates listed key attributes by which we can achieve success without victory: “Are our interests protected? Is the security of the United States protected? Are the Americans safer at the end because of the sacrifice these soldiers have made? That’s the real question.” 26

Before he replaced Robert Gates as Secretary of Defense, Leon Panetta was the Director of Central Intelligence. In June, then CIA Director Panetta said, “The fundamental mission is to provide sufficient stability so [that nation] is never again a safe haven for al Qaeda and similar groups.”27

Secretary of State Hillary Clinton has defined the terms of success in Afghanistan in several ways over the last two years. In May of last year, she focused on efforts to reconcile the Taliban with the Afghan government. “For reconciliation to succeed,” she told reporters after the raid that killed Osama bin Laden, “Pakistan must be part of this process.”28 In 2010, Secretary Clinton told Greta van Susteren in Afghanistan, “a secure and stable Afghanistan [means] a country that is able to defend itself with an army and a police force that is able to protect its citizens from the kind of terrorist activities that now plague many parts of the country.” 29

General David Petraeus, speaking during his testimony before the Senate Armed Services Committee in March of this year, said, “it’s important to remember why we are [in Afghanistan]. That’s where 9 /11 began — that’s where the plan was made... We do see al Qaeda looking for sanctuaries all the time.”30

President Obama’s first Special Representative for Afghanistan and Pakistan, Ambassador Richard Holbrooke, for his part, told the Senate Foreign Relations Committee in July of 2010, that when he references an “end-state,” the definition of success, it should be “a sustainable end-state which includes continued American economic and development assistance.”31

During his most recent trip to Afghanistan, President Obama signed a strategic partnership agreement with the Afghan government, signaling a shift in bilateral relations. While the war’s overall objectives have not changed, President Obama emphasized that the current goal of U.S. forces in the country is to successfully “transition” primary responsibility for Afghanistan’s security from international forces to local ones – thus allowing for a withdrawal of combat troops.32

Taken together, these senior officials reveal a worrying ambiguity about the war’s strategic goals and desired end-state. The Afghan government is far from stable or capable of exercising sovereignty over its territory; the insurgency remains undaunted from launching horrifying attacks into the biggest cities; civilian casualties in the conflict are higher than they have ever been.33

Even more troubling, they define victory in the war as absence: Afghanistan is a success if X does not happen. It is an impossible goal, as a single act could lead to defeat. By President Obama’s definition of victory, success means al Qaeda can never use Afghanistan or Pakistan to threaten the United States.

This then means the war is either already won, or it will never be won—al Qaeda does not have a safe haven in Afghanistan and is “on the ropes,”34 or, alternatively, we must stay there forever to make sure it never has a safe haven in Afghanistan or Pakistan.

This is strategic incoherence at its most stark.

Despite the vagueness of these goals for victory, we can look at certain indicators to determine whether the overall environment in Afghanistan is improving or degrading. By doing so, the likelihood of victory, or the specter of defeat, becomes apparent.

The Metrics

At a conference in 2009, the late Ambassador Holbrooke was asked how he would measure progress in Afghanistan. Holbrooke responded, “In the simplest sense, the Supreme Court test for another issue: We’ll know it when we see it.” 35

While many pundits and analysts cringed at the reference (Supreme Court Justice Potter Stewart, in discussing his definition of pornography in 1964, said, “I know it when I see it.” 36), Holbrooke revealed a common problem in measuring the war’s progress: it is often not clear what we should measure, and how to interpret the data.

We broke down President Obama’s strategic objectives for the war into three broad pillars, and from that derived nine broad metrics that should tell us whether or not the war is being won. The metrics we discuss below are not meant to capture the entirety of the war, or even necessarily to capture a representative sample of it. Rather, these nine metrics show how the war as a system, and not just a few points of data, is progressing.

The three pillars of President Obama’s strategic objectives requires some explanation:

  1. Deny al Qaeda safe haven. The term “safe haven” is problematic, as it can mean many things. There have been many arrests and trials for attempted terrorism in the U.S.37, but very few analysts would say the U.S. is a “safe haven” for terrorism. We chose to interpret this statement to mean what it meant when discussing al Qaeda in Afghanistan in the 1990s: large training camps where recruits are indoctrinated, instructed in how to conduct terrorist operations, and sent abroad to do so.38

  2. Prevent the Taliban from overthrowing the government. President Obama did not say “defeat the Taliban,” or “end all violent attacks,” he said he aimed to “reverse their momentum” and prevent the government’s overthrow. The two key words we use to derive metrics are “overthrow” and “government.” There is a difference between a state and a government — a state is an organized political community occupying a certain territory, while a government is an organization that has the power to make and enforce laws on that territory. The insurgency is fighting for the Islamic State of Afghanistan39 — a fundamentally different state than Hamid Karzai’s regime currently in control of Kabul. In order to build that different state, they must overthrow the current government — that is, they must destroy or displace the current regime, and take over the levers of government to create their alternative state. It is a social and political act as much as it is a physical one.

  3. Build up the Afghan security forces and the government so they can take responsibility for their country’s future. Both governments and security forces have basic attributes to be able to rule their territory effectively. Governments need sovereignty, legitimacy, and the rule of law in order to govern. Security services need to have a stable cadre of members trained so that institutional practices can be developed and a sense of professionalism can prevent widespread abuses. Neither of these two is particularly easy to quantify and measure, much less “build up,” so we tried to pick metrics that would indicate how both sets of institutions are behaving and being perceived.

Using these three pillars, we derived nine metrics that will indicate whether those three objectives are being met:

  • Political Participation helps us understand if the Afghan government has sufficient capacity to “take responsibility for Afghanistan’s future.” If Afghans feel comfortable participating in the political process, then they feel confident in their government’s ability to function. Conversely, if Afghans either feel excluded from political life, or feel like their participation is meaningless, then they do not believe their government has the ability to function.

  • Violent Rhetoric from Religious and Community Leaders helps us understand if the insurgency’s momentum is really being reversed, and if there is sufficient public sentiment to support their overthrow of the Afghan government. If religious and community leaders are constantly calling for violence against Afghan government and ISAF forces, that indicates deep-seated resentment against the mission.

  • “Shadow” Government and Judicial Institutions are a good indicator of where the Taliban and other insurgent groups feel most comfortable operating. By measuring the prevalence and functionality of these insurgent institutions we can get a sense of where government control is being superseded by the insurgency, which will tell us where government capacity is being most threatened.

  • Provincial and Local Institutions and Accountability tells us how Afghans feel about the local institutions of government. If Afghans are participating in and engaged with their local government, they are less likely to support an insurgent overthrow of the government. This metric also indicates how Afghans perceive the state’s ability to function. While some informal institutions, such as tribes or village councils might appear appropriate substitutions, President Obama’s strategy requires a government, so we focus in this metric on presence and performance of the government.

  • Agricultural Production helps us measure the economic foundation of the country. If staple crops are being grown in an economically sustainable way, then that will form the foundation of a stable government that can control its territory. It also indirectly demonstrates local security conditions, as well as the state of infrastructure for moving goods to market.

  • Afghan National Security Forces Retention is a crucial factor in determining the long-term prospects of the ANSF. If there is a high retention rate that is effectively a vote of confidence in both the security forces and the government they serve. If retention is low, that indicates low confidence in the prospects of success for the security services and the government. In addition, the rise of “green-on-blue” attacks, where Afghan forces attack and kill their western mentors, is a growing concern as well.

  • Childhood Literacy is a leading indicator of the long-term prospects of a stable Afghan government. A growing rate of childhood literacy will make extremism less likely to take root in Afghan communities. Literacy is also a strong indicator of the prospects for the development of civil institutions, which are a crucial component of a stable government.40

  • Self-Identified Membership in al Qaeda is the first objective President Obama routinely states in his strategic commentary on the war. One of the primary goals of the war in Afghanistan is to cripple al Qaeda and deny them safe haven or the ability to launch attacks. While self-identified membership is not directly related to their capacity for launching attacks, changes in al Qaeda’s self-identified membership over time is an indirect indicator of how healthy the organization is, and of the pool of people it has access to inside Afghanistan in order to launch attacks.

  • Violence is vital to understanding the war. If violence is increasing over time, it may or may not indicate mission failure: the nature of the violence, including who perpetrates it and against whom it is perpetrated, is paramount. A granular understanding of the system of violence in Afghanistan will let us understand if the war is working toward the future Afghanistan controlled by a stable government, stable security services, and free of the existential threat from the Taliban or from Al Qaeda using its territory to launch attacks. We do not try to capture in this paper the specific numbers of any particular metric. Rather, we aim to capture whether or not these metrics are being measured, and just as importantly if we can track how those metrics are changing over time: is it improving, is it getting worse, and what implications we can derive from that metric’s trajectory.

These indicators are in no particular order, but the data they provide should be judged as a group, and not solo.

No one indicator stands out on its own as being solely indicative of success or failure—the war is a system.

Political Participation

An increase of non-violent political involvement should be seen as a metric of success. This should include both formal and informal political networks as well as movements that are pro- and anti- Afghan government. If the Taliban, for example, decides it would rather oppose the Karzai government through running for office and holding protest marches, that is a far better end-state than continued fighting.

More non-violent anti-government involvement indicates opposition forces turning away from violence and becoming part of the political framework. This is a good thing, as the non-violent, political resolution of disputes is an important milestone in ending the uncontrolled violence.

There are extensive data for electoral participation. Several agencies compile data on voter turnout for Afghanistan’s elections. The International Institute for Democracy and Electoral Assistance, for example, has national-level data for both Parliamentary and Presidential elections from 2004-2010.41 The Free Elections Foundation of Afghanistan catalogues Afghanistan’s voter registration process,42 and created a detailed map of polling centers, voting irregularity and other incidents of fraud and intimidation, Election Day violence, and other voting issues.43 The National Democratic Institute created a repository of freely available data about the Parliamentary and Presidential elections from 2004-2010.44 The International Republican Institute also compiled detailed reports on electoral participation in both the Parliamentary 45 and Presidential46 elections.

However, there are many data that would also lend insight into political participation that we could not find:

  • Numbers attending political meetings, organizations, march, protests etc (male and female)
  • Number of non-violent protests
  • Number attending shura/jirga/council meetings (male and female)
  • Number of individual religious and community leaders calling for political involvement (by frequency, region, and affiliation)

There are many others. These are the basic building blocks of a civil society, which is normally expressed through participation in elections.

The elections process, too, is deeply flawed – while the reports many NGOs assemble are unequivocal about the problems that result from such a flawed election system, a look only at numbers of voters over time may not reveal if Afghans are actually buying into the idea of participating in the political system or not.

As it stands now, we could not assemble sufficient data to tell whether Afghans are becoming more or less involved in political issues.

Violent Rhetoric from Religious and Community Leaders

In contrast to non-political involvement, leaders calling for violence, against any party or institution should be seen as a negative. Preaching or calling for violence indicates an extreme rejection of the state.

Violent rhetoric could potentially be important, if only as an early indicator of areas that may be prone to violence later. In May of 2012, preacher Terry Jones held a public rally in Florida where he lit a Koran on fire. While the event was largely unnoticed in the U.S., in Afghanistan several mullahs in the northern city of Mazar-i Sharif preached that Muslim worshippers must exact revenge for the desecration of their holy book.47 In the ensuing riot, protesters angry at the lack of Americans to attack nearby stormed the local UNAMA compound, breaching its walls and killing seven U.N. workers. Rioters stirred to action by violent preaching by religious leaders in Kandahar killed another ten people that same day.48

In February, the accidental burning of some Qurans at Bagram Airfield prompted days of occasionally violent protests, at least some of which were reportedly whipped up by firebrand preachers.49

After extensive searching, we could not find a single comprehensive study about the nature, extent, or content of the rhetoric used by religious or community leaders.

Shadow Government and Judicial Institutions

This is not a measurement of informal local government or justice institutions, but rather institutions created by the insurgency to contest or replace state control. Shadow institutions represent a coercive impact on the state.

As the U.S. and its allies try to assist the Afghans in building an accountable, responsible government, a growth in the number and extent of these shadow institutions would represent a failure. At the same time, however, a decrease in these structures does not necessarily represent a victory of the government – such a decrease would simply mean the insurgency is not replacing those that are destroyed.

Much like religious and community leader rhetoric, there are little data about Taliban institutions. While this is for understandable reasons owing to the difficulty of measuring, the paucity of data – limited in some cases to mere anecdotes – makes drawing conclusions about shadow institutions difficult.

We could find no comprehensive study of the number of and type of shadow institutions, or even a reliable list of known shadow officials (governors, judges, and so on). The Afghanistan Research and Evaluation Unit completed a case study in 2010 about governance structures in Nimroz Province, but it relied on anecdotes and interviews rather than hard data.50 While The Brookings Institution presents data about how and where Afghans choose to pursue justice, their data is from 2007.51

The data on shadow institutions is badly incomplete. No one is measuring it, so it is impossible to draw conclusions from it.

Provincial and Local Institutions, and Accountability

Related to the issue of shadow institutions, success in Afghanistan as defined by President Obama must involve Afghans having an accountable local government by and for Afghans. There are more data about local government performance than there are about shadow institutions, but the data are still incomplete.

As one example, we could not find any data of how many government institutions have literate officials and staffs. This is a critical issue in measuring government performance, as if officials and staff cannot read they cannot participate in a modern bureaucracy. In fact, the only data we could find about the performance of Afghanistan’s governmental institutions relate to failures. A 2011 Congressional Research Service report catalogued how local governments in Afghanistan fall short of their goals and listed how many positions remain unfilled.52 AREU has written a number of reports about government failures,53 but these are based mostly on anecdotes and not hard data.

We also do not have data on the makeup of local governance institutions: how many officials are elected by their communities, how many are appointed from Kabul, and how many are placed by Coalition forces. We do not have sufficient data to measure the functionality and accountability of local governments in Afghanistan.

Agricultural Production

Agricultural production (licit or illicit) is linked to the use of land and the amount of capital and time invested. If farmers are growing more food on more land, that indirectly indicates improvements in security and infrastructure, along with an improvement in stability. Farmers will feel more comfortable growing more food if they have a stable economy and stable politics.

The mixture of licit and illicit agriculture also matters: if farmers tend to grow more cannabis or opium than they do licit crops, that also indirectly indicates the relative level of government control, presence of organized criminal networks, and overall confidence in the government.

The challenge with measuring agricultural production in Afghanistan, however, is that the few indices available aren’t always specific in their methodologies. So when IndexMundi compiles data from the USDA, there is no indication of how it was originally gathered. Moreover, given the shape of the graph above, which tries to chart Afghanistan’s cotton production, the long flat lines probably indicate little or no data (it is uncertain which). The data on cotton production are probably unreliable, as are IndexMundi’s charting of rice production.54 They draw their data for both charts from the United States Department of Agriculture, but there is no indication on how the data are collected or organized. The USDA also publishes data on several commodities in Afghanistan like wheat.55 All datasets show enormous variations during the 2001-2010 period, unless there are little or no data (like the long flat lines for cotton production). It is difficult to draw conclusions from this data.

There are relatively more data for opium production. The United Nations Office on Drugs and Crime assembles yearly reports about the production of opium in Afghanistan. They have data from 2005-2011, and show a very mixed record: changes to opium cultivation in recent years seem to be the result of weather and food price fluctuations as much as anything else.56

All told, agricultural production is one of the most closely-tracked metrics. Despite that, the data one might use to draw conclusions is deeply flawed, and does not seem to obey clear trend patterns useful for analysis.

ANSF Retention

President Obama’s victory definition requires a strong Afghan security sector, most commonly referred to as the “Afghan National Security Forces.” Gauging the effectiveness of the ANSF can take many forms: assessments of operational independence,57 an analysis of police corruption that fuels resentment toward the government,58 measurements of brutality against Afghan citizens,59 or even how often those Afghan forces turn violent against their own.60

We think focusing on retention is a more neutral measurement of ANSF capability, though it is concerning to see the rise of Afghan-on-Coalition killing. A high retention rate in ANSF organizations represents a general belief in victory, as well as demonstrating confidence in the opportunity of service and lessening violence. A rise in “green-on-blue” violence indicates either insurgent infiltration of the security services or a catastrophic loss in confidence on the part of the Afghans (or both).

There are fairly good data on ANSF retention: according to a recent report, desertion rates jumped 46% in 2011.61 That is, more and more ANSF personnel are walking away from their service duties as compared to previous years.

There is also the rise of green-on-blue attacks. Since 2003, 83 Nato deaths have been at the hands of their Afghan colleagues.62 Afghan soldiers killed 34 Coalition troops in 2011; in 2012 at least ten separate attacks have killed at least 17 NATO soldiers. The military is also facing allegations that it is consistently underreporting attacks against NATO forces by Afghan security forces.63

Both trends indicate an overall breakdown of confidence in the ANSF mission.

Childhood Literacy

Most studies of education in Afghanistan focus on the number of schools built, or the number of children with access to education, or the number of children who attend school. These are all interesting measurements, but they don’t capture what is actually happening at those schools.

It is important that children not only attend school, but learn at them. Thus, measuring childhood literacy, and not just school attendance, captures the effect of education programs, and not just their existence.

Furthermore, a growing number of literate children indirectly demonstrates several phenomena: the belief and trust in education and the state, the ability of local populations to send their children to a school, and that schools are adequately teaching those children.

The Brookings Institution compiled several studies to measure annual enrollment in elementary secondary education from 2002-2011.64 While the trend lines are positive, this metric does not capture literacy rates. UNICEF claims a 49% literacy rate in Afghans aged 15-24, but these data are sourced to an unknown UNESCO survey.65 The most recent UNESCO country programming document, from 2010-2011, uses data from the Afghanistan National Vulnerability Risk Assessment report, published by the Ministry of Rural Rehabilitation and Development. It estimates literacy in Afghans above the age of 15 to be 26% (39% for men and 12% for women).66 It is unclear how these statistics can be reconciled, but most surveys seem to agree that both access to and quality of education is improving.

Self-Identified Members of Al Qaeda

The overwhelming reason U.S. citizens are risking their lives in Afghanistan is to degrade and to deny territory to al Qaeda — as stipulated in President Obama’s stated strategic rationale for the war. Is the number of al Qaeda in Afghanistan growing or shrinking? Lowering the number of al Qaeda is a key measurement of mission success.

We distinguish between insurgents described by U.S. officials as being “tied to al Qaeda” and actual, self- identified members of al Qaeda. It is unclear how the U.S. government distinguishes between associations with al Qaeda from membership in al Qaeda, which makes determining the effect of such a measurement an inappropriate basis for analytic conclusions. As one example, in early September 2011, the Coalition killed Sabir Lal, a man previously incarcerated in Guantanamo.67 Officials said he was responsible for organizing attacks in the volatile Kunar province, and that he had “links to al Qaeda.” However, the nature of those links, and what it really means to be “linked” to al Qaeda, was left unstated apart from his being “in contact with senior al Qaeda figures in Kunar and Pakistan.” We have no idea what the designation means in practice: was Melma taking orders from these senior al Qaeda figures, or was he merely coordinating with them because they operated in the same area? Answering that question is critical to understanding the real nature of al Qaeda, and whether Coalition efforts to degrade its capabilities are working.

There are very few data to indicate the number of self-identifying al Qaeda in Afghanistan. In 2010, CIA Director Leon Panetta said, “at most, we’re looking at maybe 50 to 100 [members of al Qaeda in Afghanistan], maybe less.”68 It is unclear if he meant self-identified al Qaeda or those the U.S. identifies as having links with al Qaeda. In reporting on the September 2011 Sabir Lal killing, ISAF claimed to have “captured or killed more than 40 al-Qaida [sic] insurgents in eastern Afghanistan this year.” 69 However, an analysis of ISAF reports on al Qaeda shows that they only reported 22 killed and 10 captured during 2011.70

We do not have sufficient data to show for certain what the state of al Qaeda in Afghanistan really is, or how it is changing. If ISAF really has captured or killed 40 a Qaeda fighters in 2011, we still do not know how that affects Secretary Panetta’s 2010 estimate of “50 to 100” al Qaeda members fifteen months ago.


By far the most common metric applied to Afghanistan is a measurement of violence. Methodologies for measuring violence vary tremendously, and official agencies disagree about how to observe it. In September, both ISAF and the UN released conflicting reports of violence in Afghanistan: ISAF argued there is less violence in the country, while the UN claimed there was substantially more. 71

The difference between ISAF’s violence metric and the UN’s violence metric is most likely definitional: ISAF, for example, does not consider the planting of an IED to be a violent act if it is defused and no one is injured.72 The UN, on the other hand, uses a more traditional measurement of violence, in which the very act of planting an improvised explosive, whatever the result, is an act of violence against the state.

Most measurements of violence in Afghanistan are too broad for drawing analytic conclusions. Less violence is generally a good thing, though it does not actually mean the Taliban is being defeated; more violence doesn’t automatically indicate failure, either, it only means government control is being contested. Hence, measuring violence can require granularity depending on the focus of measurement.

A concentration of violence against civilians probably means something different than a concentration of violence against Afghan government officials or ISAF troops.

Different sources compile their data differently, with no set definition of what constitutes combatants and non-combatants, and, as with ISAF’s statement above, no set definition of what constitutes violence.73 As such, we do not believe aggregate numbers about violence say very much about the state of the war. Rather, violence in Afghanistan is part of a social and political system, and should be viewed in that context.

In terms of the social and political effects of violence, the last two years in Afghanistan have been devastating.

While attacks like September’s brief siege of the U.S. Embassy in Kabul did not do substantial damage or kill more than a few people, the social and political consequences of that attack are quite significant. As a symbol, attacking the U.S. Embassy is a devastating statement on America’s stature in Afghanistan, on the efficacy of ISAF and Afghan security forces in Kabul, and on the insurgency’s belief in its chances for victory. Rather than being “harassment,” as U.S. Ambassador Ryan Crocker described the attack,74 it instead punctured the widespread Afghan belief than American diplomats in their embassy were untouchable.

At the same time, ISAF officials are correct in a way to describe such actions as being relatively minor and of little consequence to anyone’s ability to fight.

Attacking the U.S. embassy, or an international hotel, or even murdering an Afghan government official does not alter the physical status of the war – it does not change the number of U.S. troops, or their ability to move across the countryside and occupy territory, or the ability of the Afghan government to function. From a physical measurement of war, which ISAF relies on to gauge its success, these attacks, however complex and “spectacular,” 75 do not alter the war in any significant way.

The problem of violence is a perfect encapsulation of the Two Wars concept we explain above: the Taliban are not fighting a physical war, but a political and social war; ISAF, on the other hand, is fighting a largely physical war and does not measure the political and social aspects of the war.

The symbolic, low-damage, low-casualty attacks in Kabul have not had much of a physical effect on the war, but their political and social consequences have been enormous, and Afghanistan is currently struggling with a sharp “brain drain,” of educated Afghans fleeing the country to live and work away from the chaos.76

Insurgent violence, like planting an IED or laying siege to a building owned by foreigners, matters because it is an act of political violence: it is an attack on the state, a manifestation of someone rejecting participation in the political process. In a similar vein, non-insurgent violence, like police brutality and criminal violence are only relevant to U.S. strategic decision-making in terms of how they alter the political participation of ordinary Afghans.

It is the political and social consequences of violence that matter to the question of whether we are winning the war — not the violence itself.

There are no comprehensive studies of the political and social consequences of violence. In other words, we are not measuring violence in a way that will help us understand if we are progressing toward President Obama’s three objectives or not.

If we are to consider violence a political and social act, then we should try to measure the political and social effects of violence.

These metrics are mostly correlative: if an attack or series of attacks are accompanied by a rapid change in a metric like political participation, religious and community leader rhetoric, or insurgent shadow government institutions, then we can start piecing together a picture by which violence affects Afghanistan’s political and social climate.

In a political war, violence only matters in terms of how closely it correlates to sudden changes to the status quo, or how a persistent pattern of violence happens at the same time as other trends.


In 2012, over 10 years after its start, it has become clear that there are two wars in Afghanistan — a physical one being fought by ISAF and a war of influence being fought by insurgents.

President Obama’s strategic objectives in Afghanistan are undefined and by their very nature — unachievable. They do though, imply a set of phenomena that can be tracked. We attempted to derive metrics, and to catalogue whether they are being measured.

There are no public data that capture these metrics in their entirety, and many metrics we should be attempting to measure are simply not tracked in publicly available data.

What is clear, though, is that we are not tracking our progress in the war.

The three pillars of President Obama’s strategy are:

  1. Deny al Qaeda safe haven.
  2. Prevent the Taliban from overthrowing the government.
  3. Build up the Afghan security forces and the government so they can take responsibility for their country’s future.

Using these three pillars, we derived nine metrics that will indicate whether those three objectives are being met:

  1. Political Participation
  2. Violent Rhetoric from Religious and Community Leaders
  3. “Shadow” Government and Judicial
  4. Provincial and Local Institutions and Accountability
  5. Agricultural Production
  6. Afghan National Security Forces Retention
  7. Childhood Literacy
  8. Self-Identified Membership in al Qaeda
  9. Violence

Several success metrics are not tracked at all; the rest are tracked incompletely at best.

Because the strategy lacks specificity, direction, and an end state in the strategy, it is easy for analysts, pundits, and policymakers to project their own assumptions onto news about the war.

Without a solid grasp of the data that demonstrate success or failure as laid out by the metrics in this paper, there is no way to know if the war is being won or not.


  1. Fox News, “Afghan Police: 8 Die in Kabul Supermarket Blast,” January 27, 2011, ↩︎

  2. Shalizi, Hamid and Robinson, Matt, “Suicide Attack on Kabul Hotel Kills Two,” Reuters, February 14, 2011, ↩︎

  3. Shalizi, Hamid and Graham-Harrison, Emma, “Suidide Attack Hits Afghan Army Bus in Kabul: Police,” Reuters, April 9, 2011, ↩︎

  4. CNN, “Attack on Kabul Hotel Ends with Deaths of 7 Taliban, 11 Others,” June 28, 2011, ↩︎

  5. Associated Press, “Afghan Officials: String of Attacks Kills 21,” June 12, 2011, ↩︎

  6. Associated Press, “Deadly Explosions Rock British Council Compound in Kabul,” August 19, 2011, ↩︎

  7. Rubin, Alissa J.; Rivera, Ray and Healy, Jack, “US Embassy and NATO Headquarters Attacked in Kabul,” The New York Times, September 13, 2011, ↩︎

  8. Ackerman, Spencer, “Taliban Kill 13 Americans in Armored Bus Attack,” Wired Danger Room, October 29, 2011, ↩︎

  9. Nissenbaum, Dion, “Taliban Attacks Kill 13, Reflect Strategic Shift,” The Wall Street Journal, December 30, 2011, ↩︎

  10. “Taliban kill 20 Afghans in suicide attack, bombings in the south,” The Long War Journal, January 19, 2012, ↩︎

  11. Nissenbaum, Dion, and Khan Totakhil, “Afghan Insurgents Strike Across Nation,” The Wall Street Journal, April 16, 2012, ↩︎

  12. Shah, Amir, and Chris Blake, “Taliban kill 7 in Kabul after Obama visit,” The Associated Press, May 2, 2012, ↩︎

  13. Graham Bowley, “Taliban Kill 5 Afghan Education Officials Near Border,” The New York Times, May 8, 2012, ↩︎

  14. Foust, Joshua, “The Unbearable Lightness of Kabul,” Columbia Journalism Review, December 1, 2008, ↩︎

  15. Exum, Andrew, “What We Still Don’t Know about the Kabul Attacks,” Foreign Policy, September 13, 2011, ↩︎

  16. Wikipedia contributers, “Ring of Steel (Kabul),” Wikipedia, The Free Encyclopedia, ↩︎

  17. Ackerman, Spencer, “Team Petraeus Brings Body Counts Back,” Wired, March 8, 2011, ↩︎

  18. Smock, Dan, “ISAF Reporting and the IED Numbers Shuffle,”, October 17, 2011 ↩︎

  19. Clark, James W., “Afghan Officials, Elders Hold Shura in Marjah,” May 22, 2010, ↩︎

  20. Michaels, Jim, “General: Heart of Afghanistan Insurgency Beaten,” USAToday, February 14, 2011, ↩︎

  21. Obama, Barack H., “Remarks by the President in Address to the Nation on the Way Forward in Afghanistan and Pakistan,” The White House, December 1, 2009,↩︎

  22. Wong, Kristina, “Obama Promises to Bring Troops Home from Afghanistan Beginning July 2011,” ABC News, January 26, 2011, ↩︎

  23. Bennett, John T., “White House Redefines ‘Winning’ in Afghanistan Ahead of Review,” The Hill, June 6, 2011, ↩︎

  24. CBC News, “U.S. Can’t Fail Afghanistan, Obama Nominee Warns,” June 8, 2011, ↩︎

  25. Bennett, John T., “White House Redefines ‘Winning’ in Afghanistan Ahead of Review,” The Hill, June 6, 2011, ↩︎

  26. Ibid. ↩︎

  27. Bennett, John T., “Panetta has Tough Talk for US Foes, Defines Afghanistan Mission,” The Hill, June 9, 2011, ↩︎

  28. Associated Press, “US-Pakistan relations ‘at turning point’ after Killing of Bin Laden, Warns Clinton,” May 27, 2011, ↩︎

  29. “Interview with Hillary Clinton,” Fox on the Record with Greta van Susteren, July 20, 2010, ↩︎

  30. Crabtree, Susan, “Petraeus on Afghan War: ‘We Have To Remember Why We’re There’,” Talking Points Memo, March 15, 2011, ↩︎

  31. Gaouette, Nicole, “Obama Must Define Afghan War Goals, Senators Tell Holbrooke,” Bloomberg, July 14, 2010, ↩︎

  32. Landler, Mark, “Obama Signs Pact in Kabul, Turning Page in Afghan War,” The New York Times, May 1, 2012, ↩︎

  33. Haidary, Emal, “Civilian deaths threaten US pace: Afghan president,” AFP, May 8, 2012, ↩︎

  34. Dozier, Kimberly, “US Counterterror Chief: Al-Qaida Now on the Ropes,” Associated Press, September 1, 2011, ↩︎

  35. Tiedemann, Katherine, “Holbrooke on Success: We’ll Know it when we See it,” Foreign Policy, August 12, 2009, ↩︎

  36. Wikipedia contributors, “I Know it When I See it,” Wikipedia, The Free Encyclopedia, ↩︎

  37. CNN, “Spate of Terrorism Arrests not Connected, Analysts Say,” September 25, 2009, ↩︎

  38. Wikipedia, ↩︎

  39. Rediff, “Taliban wants an Islamic state in Afghanistan: Report,“November 3, 2008, ↩︎

  40. Curtis, Lisa, “U.S. Aid to Pakistan: Countering Extremism Through Education Reform” The Heritage Foundation, June 8, 2007, ↩︎

  41. International Institute for Democracy and Electoral Assistance, “Voter Turnout Data for Afghanistan,” November 8, 2010, ↩︎

  42. Free & Fair Election Foundation of Afghanistan, “2009 Voter Registration Observation,” May 2009, ↩︎

  43. Free & Fair Election Foundation of Afghanistan, “Afghan Election Mapper,” September 18, 2010, ↩︎

  44. National Democratic Institute, “Afghanistan Election Data,” ↩︎

  45. The International Republican Institute, “Afghanistan Parliamentary and Provincial Council Elections, September 18, 2005: Election Observation Mission Final Report,” ↩︎

  46. The International Republican Institute, “Afghanistan Presidential and Provincial Council Elections, August 20, 2009: Election Observation Mission,”↩︎

  47. Najafizada, Enayat and Nordland, Rod, “Afghans Avenge Florida Koran Burning, Killing 12,” The New York Times, April 1, 2011, ↩︎

  48. BBC News, “Afghanistan: Deadly Kandahar protest at Koran burning,” April 2, 2011, ↩︎

  49. van Bijlert, Martine, “The Quran burnings and the different faces of anger,” Afghanistan Analysts Network, February 22, 2012, ↩︎

  50. Larson, Anna, “Governance Structures in Nimroz Province,” Afghanistan Research and Evaluation Unit, November, 2010, ↩︎

  51. Livingston, Ian S.; Messera, Heather L. and O’Hanlon, Michael, “Afghanistan Index: Tracking Variables of Reconstruction & Security in Post-9 /11 Afghanistan,” The Brookings Institution, May 31, 2011, p. 22, ↩︎

  52. Katzman, Kenneth, “Afghanistan: Politics, Elections, and Government Performance,” Congressional Research Service, pp. 33-55, ↩︎

  53. For example, see: Coburn, Noah, “Losing Legitimacy? Some Afghan Views on the Government, the International Community, and the 2009 Elections,“Afghanistan Research and Evaluation Unit, November\ 2009,,%20the%20International%20Community,%20and%20the%202009%20Elections%20Bf%202009.pdf ↩︎

  54. Index Mundi, “Afghan Milled Rice Production by Year,” 2011, ↩︎

  55. U.S. Department of Agriculture, “Afghanistan: Wheat Production Recovers in 2009/10 Season,” June 10, 2009, ↩︎

  56. U.N. Office on Drugs and Crime, “Afghanistan Opium Survey 2011,“January 2011, p. 26, ↩︎

  57. Ackerman, Spencer, “Not a Single Afghan Battalion Fights Without U.S. Help,” Wired, September 26, 2011, ↩︎

  58. Harding, Thomas, “Afghanistan Police Corruption is Fuelling Insurgency,” The Telegraph, June 3, 2010, ↩︎

  59. Healy, Jack, “Brutality by Local Afghan Police is Reported,” The New York Times, September 12, 2011, ↩︎

  60. Walsh, Nick Patton, and Masoud Popalzai, “2 American officers killed in Afghan ministry,” CNN, February 25, 2012, ↩︎

  61. Partlow, Joshua, “One in Seven Afghan Soldiers Deserted During First Half of Year,” The Washington Post, September 2, 2011, ↩︎

  62. New America Foundation, “Attacks on US and NATO Soldiers by Afghan Security Forces,” ↩︎

  63. Burns, Robert, “US Not Reporting All Attacks by ‘Friendlies,’” The Associated Press, May 1, 2012, ↩︎

  64. Livingston, Ian S.; Messera, Heather L. and O’Hanlon, Michael, “Afghanistan Index: Tracking Variables of Reconstruction & Security in Post-9 /11 Afghanistan,” The Brookings Institution, May 31, 2011, p. 31, ↩︎

  65. UNICEF, “Afghanistan Statistics,” March 2, 2010, ↩︎

  66. UNESCO, “UNESCO Country Programming Document (UCPD): Afghanistan,” 2010-2011, p. 10, ↩︎

  67. CNN, “NATO: Man with al Qaeda links killed in Afghanistan,” September 3, 2011, ↩︎

  68. Sonmez, Felicia and DeLong, Matt, “Panetta: 50-100 al-Qaeda Remain in Afghanistan; Cornyn: GOP is the Party of Change; Leahy, Sessions Preview Kagan Confirmation Hearings,” The Washington Post, June 27, 2010, ↩︎

  69. ISAF Joint Command, “ISAF Joint Command Morning Operational Update Sept 3, 2011,” September 3, 2011, ↩︎

  70. Strick, Alex, “Those ‘40 al-Qaeda insurgents’...,” A Different Place, September 13, 2011, ↩︎

  71. Nissenbaum, Dion, “Differing Data Shape Views of Surge,” The Wall Street Journal, September 6, 2011, ↩︎

  72. ISAFmedia, “Twitter Post @joshuafoust,” September 29, 2011,!/ISAFmedia/status/119381386514800640 ↩︎

  73. We found data compiled by The Guardian for 2007-2010: The Guardian, “Afghanistan Civilian Casualties: Year by Year, Month by Month,” March 10, 2010,, and The United Nations Security Council publishes quarterly reports for the Secretary-General: U.N. Security Council, “The Situation in Afghanistan and its Implications for International Peace and Security,” September 14, 2010, pp. 12-13, The Afghanistan Conflict Monitor assembled datasets of Civilian Deaths by Perpetrator and District: Human Security Report Project, “Afghanistan Conflict Monitor: Civilians,” March 9, 2011,, as well as “Armed Opposition Attacks by Province and Year, 2008-2010”: Human Security Report Project, “Afghanistan Conflict Monitor: Violent Incidents,” February 14, 2011, The Afghanistan NGO Safety Office has violence data for 2006-2010: The Afghanistan NGO Safety Office, “ANSO Quarterly Data Report,” April 2011, pp. 10-12, The Congressional Research Service issued a report to Congress in December of 2010 about the United Nations Assistance Mission to Afghanistan (UNAMA), which included data on civilian deaths by pro-government and anti-government forces: Margesson, Rhoda, “United Nations Assistance Mission in Afghanistan: Background and Policy Issues,” Congressional Research Service, December 27, 2010, p. 17, In addition, The Brookings Institution compiles a dataset for 2006-2011: Livingston, Ian S. and O’Hanlon, Michael, “Afghanistan Index: Tracking Variables of Reconstruction & Security in Post-9 /11 Afghanistan,” The Brookings Institution, August 31, 2011, p. 15, ↩︎

  74. Abi-Habib, Maria, “Ambassador Blames Pakistan Militants,” The Wall Street Journal, September 15, 2011, http://online.wsj.comarticle/SB10001424053111904060604576570073922364258.html ↩︎

  75. Foust, Joshua, “A New Wave of Spectacular Attacks,”, February 19, 2010, ↩︎

  76. Peter, Tom A., “Can Afghanistan hang on to its newly minted college grads?,” The Christian Science Monitor, May 26, 2011, ↩︎