Getting the Basics Wrong: Key U.S. Failures in Afghanistan

After the 9/11 attacks, which weighed on America for a generation, the U.S.-led complete military victory over the Taliban in 2001 brought an unexpected opportunity to end decades of destructive violent conflict in Afghanistan. Afghans and many Americans and other foreigners who worked on the country were aware of and strove to exploit this opportunity, which would have greatly benefited U.S. interests as well. But starting from a broadly favorable situation—the “golden moment” sometimes seen in postconflict contexts—mistaken basic decisions, lacunae, and failures, both at the outset and subsequently, ultimately led to defeat.

The recently established Afghanistan War Commission, a nonpartisan body mandated by Congress, is an essential initiative whose tasks during its three-year lifetime will be to write an official history of the nearly 20-year U.S. engagement, extract lessons learned, and distill recommendations. The work of the commission understandably will take time. Its final report will come out five years after the end of the U.S.’s military presence in Afghanistan (termed “the longest war in America’s history”). But while the situation in Afghanistan was complicated, some key lessons are straightforward. And though hindsight often is clarifying, many of the mistakes and failures were apparent when they occurred and, indeed, were called out in real time.

This article will focus on a few key areas where what went wrong was strategically determinative of the outcome. It covers U.S. approaches, decisions, and actions, though other parties—Afghan politicians and warlords; regional countries, most notably Pakistan; and the Taliban itself—contributed to U.S. failure. It starts by delving into the crucial question of “why”: Why were certain key decisions, which seemed mistaken even at the time, made? 

Behavioral Blinders and Distorted Incentives

Problematic cognitive shortcuts (“heuristics”) and biases—foundational concepts in the field of behavioral economics—permeated U.S. decision-making on Afghanistan. These shortcomings affecting decision-making could not have been wished away, but greater self-awareness could have mitigated the damage and facilitated explicit consideration of alternative choices.

The “why” of U.S. decisions on Afghanistan cannot be understood without acknowledging the deep and long shadow cast by 9/11—the worst external attack on the continental United States since the War of 1812, the worst on any U.S. territory since World War II. During and after 2001, 9/11 dominated the emotional landscape and decision-making ethos in the U.S. government, largely dictating a counterterrorism (CT)-dominated approach, which was the justification for the Afghanistan military intervention in the first place.

This is a notable example of “salience bias” related to 9/11, and the visceral U.S. response it gave rise to was a good example of the “affect heuristic”—relying on current emotions in making quick decisions. While understandable, these cognitive biases tended to drown out other perspectives, country knowledge, and lessons from earlier experience, even as recently as the previous decade—the destructive mujahideen civil war that followed the 1989 Soviet withdrawal and the subsequent rise of the Taliban—not to speak of the Soviet occupation and earlier Afghan wars, as well as global experience.

Other behavioral pathologies that affected U.S. decision-making on Afghanistan included, among other things:

Taking for granted the surprisingly quick and easy U.S.-led victory over the Taliban in 2001. This resulted in inadequate attention and resources right after that and their damaging shift to Iraq—a “satisficing,” “minimum effort,” “reconstruction and nation-building lite” approach that failed to exploit windows of opportunity.

A naïve“black and white” CT approach in the immediate aftermath of 9/11. “Either you are with us, or you are with the terrorists” was the U.S. government’s position, lumping together al-Qaeda and the Taliban as enemies. At the same time, the U.S. treated “Afghan allies,” namely leaders of Northern Alliance and warlord militias that provided boots on the ground for the U.S.-led victory, as friends who could do no wrong. This facilitated their return and entrenchment, despite the predatory depredations by many of them, which contributed to the Taliban’s rise to power in the 1990s. The lumping together of the Taliban with narco-traffickers was another example of this overly simplistic approach, leading to, among other things, ill-conceived and ineffectual air strikes against drug labs in 2017 and 2018.

Naïve expectations of support from Pakistan against the Taliban, which proved unfounded. Pakistan played a “double game,” taking actions severely inimical to U.S. success, encouraging the revival of the Taliban in the years after their 2001 defeat and providing them a durable sanctuary and associated logistical and other support. The unwillingness of Pakistan’s military to crack down on escaping Taliban elements was apparent from the start.

Reluctance and slowness to adjust strategy and approaches as the situation changed. The U.S. exhibited “decision inertia”: For years after the 2001 military victory, for example, the emphasis was on rounding up or taking down the “remnants of al-Qaeda” and the Taliban, even while the latter was reconstituting in Pakistan.

Failure to take into account strategic interactions among different actors, a basic tenet of game theory whose essence is sequential moves by different “players” who respond dynamically to each other’s actions. U.S. decision-makers apparently did not take into account that the Taliban and other key actors would respond, learn, and evolve in the face of U.S. actions, not just let themselves be a static target. For example, the Taliban would reconstitute in their Pakistan sanctuary, wait out the 2009-2011 “surge,” stop massing their forces, which could be hit by U.S. airstrikes, but infiltrate in smaller groups, etc.

Wishful thinking, especially in relation to the exit strategy. The understandable search for a way out from direct U.S. military involvement in the Afghan war, for example, led to unwarranted hope being placed on Afghan National Defense and Security Forces (ANDSF), which distorted assessments and corrupted monitoring data and metrics. There were strong pressures to show the progress needed even if it wasn’t there. Other examples include the efficacy of elections in promoting short-run political stability, and leverage vis-a-vis the Taliban in negotiations while the U.S. military was drawing down.

Belief in imported institutional and legal models not fit-for-purpose in Afghanistan. The U.S. and other countries imported a host of new laws and regulations and set up new organizations, often where acceptable laws and agencies were already in place. Transplanted institutions “checked the boxes” on paper but did not fulfill the intended functions they carried out in the countries the models were drawn from.

Rampant short-termism plaguing U.S. decision-making throughout. Political imperatives to “show results quickly,” as well as bureaucratic and career incentives, helped create decisions that were reactive to the immediate situation rather than proactive and forward-looking. Assignments of U.S. military and civilian leaders and senior staff rarely exceeded one to two years, whereas time horizons for meaningful progress in the intervention were far longer (including for sustained CT success), not to speak of the time required for individual learning. There were eight U.S. military commanders in Afghanistan during 2008-2018, almost one per year. Such frequent rotations occurred in other periods and for other military and civilian staff as well. Pressures to show accomplishments, however spurious, during each campaign period and each tour of duty were understandable but counterproductive, and any accountability was greatly diffused.

A mindset that money can solve intractable problems by spending large amounts quickly, with a focus on “burn rates.” This pathology was most evident during the 2009-2011 surge period. The money deployed “as a weapons system” far exceeded Afghanistan’s absorptive capacity and fueled waste and corruption. These very harmful side effects, though well understood at the time, did not restrain decisions on spending levels.

Pervasive metrics fetishism—“everything can be measured”—a manifestation of short-termism. Especially during the surge and in specific sectors such as counterinsurgency and Afghan security sector capabilities, pseudo-scientific quantification and frequent metrics reporting (even weekly) misled U.S. policymakers and masked problems and adverse trends. This was compounded by the fact that, in most cases, units and individuals responsible for implementation also collected and analyzed the data, creating potential conflicts of interest.

Critical Mistakes and Failures

Very limited peacekeeping. Turning from the “why” to the “what,” a striking feature of the immediate post-2001 period was tiny, nontraditional peacekeeping. The International Security Assistance Force (ISAF), totaling less than 5,000 troops in the first year, initially was authorized to operate solely in the Kabul area, with the task of protecting the new Afghan government. Only after a couple of years was ISAF allowed to gingerly expand to other parts of the country, by which time the postconflict window of opportunity had faded. Having a robust UN peacekeeping presence in a country at the end of a civil war is a success factor for avoiding resumption of conflict. This should have been the go-to option for Afghanistan in 2001, but it appears to have been a nonstarter for the U.S.—probably due to the primacy of CT and misplaced concerns that a UN or other robust peacekeeping mission might have interfered with that. 

“Reconstruction lite.” In light of the huge influx of international spending and aid starting around 2006, culminating with the 2009-2011 surge—when total civilian and military aid peaked at around 100 percent of gross domestic product, it is easy to forget that Afghanistan was initially somewhat under-resourced compared to other postconflict situations. This reflected complacency after the unexpectedly quick military victory over the Taliban; the U.S.’s initial reluctance to get involved in “nation-building” (even though some of that was essential for sustainable CT); the time required to ramp up aid; and especially the shift of attention and resources to Iraq.

Undermining nascent Afghan democracy at its birth. The Emergency Loya Jirga in June 2002 was a key stage in the process instituted by the December 2001 Bonn Agreement, intended to, with inputs from the Afghan people, select a transitional administration and its head for a two-year period until a constitution was developed and the first presidential election held. The selection process for delegates worked reasonably well and injected a degree of popular will. However, the Loya Jirga was undermined in particular by two decisions.

First, just before the Loya Jirga, the U.S. pressured former Afghan King Zahir Shah to withdraw from consideration to be head of state in the transitional government. This may have reflected various U.S. considerations, but it set a problematic precedent and fostered persistent Afghan perceptions that the U.S. was driving political developments.

Second, during the Loya Jirga Northern Alliance, leaders and other warlords who had not bothered to go through the required selection process were brought in and respectfully seated, sending a strong signal of their entrenchment and impunity, leading to disillusionment and some delegates leaving the meeting.

Critical constitutional flaws. Some of the problematic features of Afghanistan’s constitution, promulgated in 2004, undermined democracy. First, the frequent elections called for in the Afghan Constitution—including for provincial and district bodies that had no clear mandates or roles—was a significant flaw. The 3- to 5-year terms for different elections implied a 60-year cycle with elections held at intervals of less than 1.5 years on average. Even leaving out the district councils (for which elections were never held), this was far too many elections for a democracy just in the process of being established. Elections in conflict and postconflict situations can be destabilizing, so later and fewer elections were called for.

Second, the single nontransferable vote (SNTV) mechanism used in parliamentary elections, with the 34 provinces as multi-representative constituencies, exists only in a couple of places in the world. It is inimical to effective political parties and to accountability of individual parliamentarians to constituents, and it enables candidates to win with only a small number of votes (arguably making it easier for local power-holders and warlords to win seats). The SNTV method is not suitable for any country, let alone Afghanistan.

More generally, Afghanistan’s constitution provided for a very strong presidency, a relatively weak national legislature (parliament), and a highly centralized state administrative structure. While in many respects hearkening back to the monarchical constitution and the prewar centralized administrative state that coexisted with a decentralized, diverse society, the highly centralized constitutional design probably was less suitable for a country and society fractured by nearly a quarter-century of war. 

Although the U.S. government was not responsible for the design features of the Afghan Constitution, it did have some influence over it. At the time, perceiving Hamid Karzai, then head of the Afghan administration, to be weak relative to other Afghan power-holders, the U.S. preference was for a strong presidency, and the disdain for political parties on the part of Karzai and others prevailed in the choice of SNTV. While Afghan politics soon developed its own pathologies and contributed to the ultimate failure, these early choices set a problematic path forward.

Not learning from the history of insurgencies. Benefiting from the Pakistan sanctuary, the Taliban morphed into an insurgency within a few years—initially ineffective but learning and growing over time. While most insurgencies fail, those that are victorious tend to have (a) staying power (winning by surviving and outlasting the government or foreign enemy); (b) sanctuary and external support, which can make or break an insurgency; and (c) some degree of local popular support or at least acquiescence in parts of the country. These lessons are well known from historical experience—earlier in Afghanistan, in Vietnam (both French and U.S. engagements), and elsewhere, in some cases going back centuries.

By the late 2000s, it was abundantly clear that the Taliban insurgency had staying power and a reliable sanctuary in Pakistan (their headquarters and political leadership located in Quetta), as well as some level of popular support in the south. Pakistan, based on its own strategic calculations, was not going to take away the Taliban’s sanctuary or stop providing covert support. On the U.S. side, both strategic considerations regarding Pakistan (its importance as a nuclear power, among other things), and the awkward dependence on Pakistan for logistical supply routes for foreign military forces in Afghanistan, apparently ruled out direct U.S. attacks on the Taliban in Pakistan.

These considerations pointed clearly to the need to reach a deal with the Taliban, and any increase in U.S. forces and kinetic military activity might have made sense only in order to be in a position to negotiate from strength. Instead, the surge was viewed as an effort to achieve military supremacy over the Taliban and provide space for the Afghan “government-in-a-box” to establish its presence and hold on to the territory taken by the U.S. military—a striking example of wishful thinking unrelated to ground realities and Afghan government capabilities. Moreover, since the surge had a preannounced end date, the Taliban just had to wait it out, benefiting from their Pakistan sanctuary. It was no surprise at all that this is what happened, and any gains achieved by the surge were temporary.

“Gold-plated,” unsustainable Afghan security forces. Building Afghanistan’s security sector started off slowly, hindered by the May 2022 Geneva agreement “balkanizing” the security sector among different “lead donors.” Once security sector formation ramped up a few years later, the Afghan National Army was repeatedly expanded in size—from an initial authorized level of 45,000 eventually to over 200,000, along with a very large police force, so the authorized strength of the ANDSF as a whole peaked at over 350,000. This even though it was pointed out by the World Bank and others starting from 2005 that the salaries and other basic costs of the ANDSF could not be covered by Afghan domestic revenues for decades.

Even worse (since a permanent influx of U.S. aid could have covered the fiscal gap), the ANDSF’s design built in long-term dependency on foreign contractors to maintain sophisticated equipment, and key enablers were either missing or, like the Afghan Air Force, developed only very belatedly, also on an unsustainable basis. When U.S. troops exited the country, essential foreign contractors also left, accelerating the ANDSF’s collapse. There appears to have been little or no advance contingency planning for a sustainable ANDSF in the event of a complete withdrawal of U.S. military forces from the country.

Wishful thinking spurred by focus on exit. With the preannounced end date for the surge, and especially after the killing of al-Qaeda leader Osama bin Laden in May 2011, the U.S. government became engaged in the search for an exit strategy. The surge was followed by transition to ANDSF troops doing the bulk of the fighting on the ground. Thus a functional and capable ANDSF was seen as key to the exit of U.S. troops. The 2014 Afghan presidential election, involving transfer of power to a new president, also was seen as facilitating the exit strategy. The search for exit resulted in wishful thinking: about the Afghan government’s effectiveness, ANDSF capabilities, a “good” 2014 election, and the ability to negotiate a decent deal with the Taliban while in the process of withdrawing U.S. troops. Failure occurred in all of these areas, and the exit strategy was chasing but never catching up to a moving target under deteriorating conditions—with the Taliban strengthening over time.

Failure to negotiate a peace settlement with the Taliban. After the U.S.-led military victory over the Taliban in late 2001, there was a window of opportunity to make a deal with the Taliban that would have involved little more than their dignified surrender with guarantees they would not be sent to Guantanamo or the Hague. Karzai wanted to make such a deal with the Taliban, but this was ruled out by the U.S. Even if the Taliban understandably could not have been invited to the Bonn Conference, this unexploited opportunity for a highly favorable peace agreement continued to be available for the next one to two years.

The next major opportunity to negotiate a peace deal with the Taliban was during the 2009-2011 surge, when U.S. and other foreign forces were at their peak and were clearing and holding core Taliban areas. While not as favorable as in 2001-2003, it was an opportunity to negotiate from a position of strength.

Even after the surge and drawdown of U.S. and other foreign troops by 2014, there were opportunities for a negotiated deal on better terms than the 2020 Doha Agreement—which prominently included a commitment for the exit of all foreign troops within 14 months.

In this context, one of the more puzzling U.S. decisions in the entire war was to kill Taliban leader Akhtar Mansoor in a drone strike in Pakistan in May 2016. Mansoor, who succeeded Mullah Omar, had business interests and appeared to be interested in negotiating; it seems almost certain that he would have been a better negotiating partner for the U.S. than either a Taliban military leader or the religious cleric (Mullah Haibatullah Akhundzada) who took over after Mansoor was killed. This is a striking example of an important decision being made apparently without thinking through the downstream consequences. A possible rationale might have been if this drone strike signaled a belated change in U.S. policy toward the Taliban’s Pakistan sanctuary, followed by further strikes against other Taliban leaders and Taliban infrastructure there. But it soon became clear that this was just a one-off operation.

So through most of the nearly 20-year U.S. intervention in Afghanistan, the primacy of counterterrorism apparently ruled out a negotiated peace deal with the Taliban. Since peace would have been important in assuring durable counterterrorism success, this approach seems inexplicable.

Civilian aid carrying unrealistic expectations. Civilian aid was undermined by numerous pathologies. From too little initially, aid burgeoned to levels far too high to be absorbed, gave rise to an artificial, unsustainable service-driven economy, created parallel structures that undermined the Afghan government, and fueled corruption and waste. Much aid was linked to counterinsurgency and stabilization, compromising its effectiveness. At the same time, aid supported improvements in social indicators, average per-capita incomes rose rapidly during 2002-2012 (albeit unequally and on an unsustainable basis), and some government institutions were built.

Though large by international standards, civilian aid was smaller than security assistance to the ANDSF and the in-country spending by U.S. and other foreign military forces. In 2018, civilian aid was estimated by the World Bank at $3.8 billion and security assistance at $4.9 billion. Reliable data are not available for foreign military spending on their own forces in-country, but total U.S. military contracts for work in Afghanistan reached $108 billion during 2001-2021, an average of $5.5 billion per year. There was also significant local spending by other countries’ militaries. So, civilian aid was not the main contributor to waste, fraud, and corruption and associated adverse political and security impacts, nor to the macroeconomic problems resulting from total financial inflows far exceeding Afghanistan’s absorptive capacity.

However, civilian aid suffered from excessive expectations about what it could accomplish. First, the idea that service delivery and basic public goods funded by aid directly promote government legitimacy is simplistic and naïve. While these aspects are important for development, they could not make up for problematic politics, poor governance, and corruption. Second, the idea that aid conditionality leads to better governance and more positive politics was even more naïve. Third, aid supporting the Afghan budget incentivized real technocratic reforms, but these reforms were not game-changers, were subject to fits and starts, and took a long time to achieve meaningful progress—as much as 10-15 years. Fourth, during the 2009-2011 surge, expectations that huge stabilization assistance would help make the military progress achieved in the short run sustainable by “winning hearts and minds” were highly unrealistic. 

***

Had greater awareness—of the Afghanistan context and historical experience there and elsewhere—and greater self-awareness regarding post-9/11 blinders fed into top U.S. policymakers’ decision-making process on Afghanistan, things might have played out differently. And incentives favoring short-termism could have been mitigated, at the very least by longer tours of duty. Behavioral and incentives-related explanatory factors go a long way toward answering questions about why the U.S. made mistakes in Afghanistan. But even taking them into account, some of the decisions seem inexplicable and will benefit from further probing into their background and processes, drawing on data the Afghanistan War Commission will have access to.