By: Ronald Neumann
The National Interest
June 6, 2025
The conventional explanations for America’s failure to stabilize Afghanistan provide little help for future policymaking.
The American memory of Afghanistan is receding in the rearview mirror. Increasingly, the potential to learn lessons from the twenty-year campaign is being wasted, replaced instead by bumper stickers and slogans that pass for knowledge but are either incorrect or largely useless without a great deal of further reflection.
Three of the most common bumper sticker lessons are “don’t do democracy,” “don’t build an army in our own image,” and “don’t do nation-building.” The problems with each of these suggest the need for deeper reflection if we are to profit from the past and get beyond slogans for future policy decisions.
Democracy in Afghanistan
The debate over how actively the United States should promote democracy abroad is nearly as old as the Republic itself. It first emerged in the early 1800s during debates over whether or how actively the United States should support liberation movements in Latin America. It is likely to continue.
The problem with using the case of Afghanistan to argue against democracy promotion as a policy goal is that it rests on the false premise that spreading democracy to Afghanistan was the principal goal of the US campaign there. In fact, the real aim throughout the Bush, Obama, and first Trump administrations was how to withdraw from Afghanistan militarily while leaving a more or less stable country behind where terrorism could not return. To do so required a basis of legitimacy on which the government could be organized. Short of returning to civil war, which had previously characterized the country, some form of peaceful allocation of power was necessary. Hence, democracy was a practical, rather than an ideological, necessity if the country was to be governed by consensus rather than bullets.
There were numerous problems in building Afghan democracy, including the time needed to establish a supporting culture and institutions, the incorrect choice of electoral system, and the difficulty of holding elections in insecure conditions. However, the problem with Afghanistan was not that democracy promotion was an unrealistic goal but rather that there were few alternatives to it.
In any case, policymakers did not frame the problem in these terms; leaving soon was a goal, but democracy was a sort of default reaction on how to achieve this. Whether that was the right choice is debatable—if one has an alternative governance model. But to conclude that the case of Afghanistan proves that the United States should refrain from democracy-building is to refuse to think about the options that were, or were not, available at the time.
No Model Army
The problems of constructing a foreign army in our own image have bedeviled US policy since the Vietnam War. Scholars have long documented how US-trained armies were not well suited to their purposes. In Vietnam, the South Vietnamese force was designed for a conventional war with the North rather than a demanding counterinsurgency.
In Afghanistan, the United States constructed a force so dependent on foreign support that it could not function without it. To take only one example, the supply system we built in Afghanistan was sophisticated, digitized, and heavily dependent on foreign expatriates, all of whom we removed at the end. However, the problem is not the truth of the slogan but rather the need for an alternative.
One cannot send large numbers of US military personnel to train the army of another country without having an organizing doctrine for training. We have no such doctrine for training a force radically different from our own, with large limitations on literacy and education. Building an army in a different model will require extensive thought and development. Without undertaking such thinking, we will be left either unable to assist in building a foreign army when one is needed or to repeat past mistakes. Thus, the phrase by itself is no help for future decisions.
State-Building, Not Nation-Building
“Don’t do nation-building” is arguably the most problematic “lesson” to emerge from recent American history. Firstly, the phrase “state-building” would be more accurate, as Afghanistan has existed as a defined state since 1747. The first Bush administration, and particularly Secretary of Defense Donald Rumsfeld, had drawn the lesson from the Balkans that nation-building was a mistake.
The result was a resistance to any commitment to institutional strengthening in Afghanistan immediately after the 2001 war when the Taliban was essentially defeated and security problems less acute. This opportune period, when foreign influence was at its peak, was largely wasted. No attention was given to building institutions. American assistance was limited to humanitarian aid only, and the first, very inadequate, developmental assistance did not begin until 2004.
In Iraq, there was an assumption that when the Saddam Hussein government was removed, Iraq would simply evolve into a democracy with very little help. The disaster of this belief has been amply documented. But the notion of not doing nation-building continued to bedevil the United States. The long-lasting second Afghanistan policy review in the Obama administration concluded that the United States would limit its goals to destroying the Taliban and not do nation-building (or state strengthening).
The problem with this formulation was that the Taliban was a regenerative movement. To keep it suppressed would require an army. But armies are part of a state, and a state needs a functioning economy and infrastructure. All of these considerations led to an enormous increase in the development budget, deployment of districts and provincial reconstruction teams, and a massive effort to increase the civilian advisory presence. The logic of these steps was unmistakable, and they unquestionably amounted to state-building, even as the administration declared it would not do so. The contradiction was not helpful to policy, to say the least.
The Real Lessons
Twenty years of warfare leave an almost endless number of decisions for debate. Tactical issues, basic governance concerns, strategies for defeating the Taliban, and shifting policies of different administrations all provide food for thought. However, there is also room to reflect on whether there are more fundamental lessons worth considering. Three recurring problems in American policymaking arise: building a learning organization, creating reasonable timelines, and identifying local partners.
Building a Learning Organization.
One important problem that is rarely, if ever, addressed is the need to build a “learning organization.” An interesting book by Georgetown professor Lise Moraj Howard compares relatively successful United Nations peacekeeping operations to search for common lessons. One lesson she drew from the successful UN missions is the need to build an institution that develops enough knowledge of the local culture and politics to implement its policies effectively.
Building a learning organization requires certain key components. The first is long-term leadership. An organization needs enough time and leadership continuity to make mistakes and climb the learning curve. In Afghanistan, the rapid turnover of ambassadors and generals, along with the deployment of a new division every year or two, was the antithesis of building a learning organization. Similarly, short tours, generally limited to one year, were also common among most military and civilian personnel.
The result was frequent changes in operational policy on the ground, in addition to the broad policy changes that came from Washington. Afghan officials grew cautious about investing too much effort in new approaches, as it was likely that, within a few months or a year, a new US official would alter the approach. When this problem is repeated over and over, it becomes increasingly difficult to get full support for any policy from the locals.
US policymakers should reconsider the length of service, particularly for generals and ambassadors, as well as the frequency of rotations for major troop units and subordinate units. Rapid rotations tend to emphasize short-term goals. Structuring organizations around the attainment of long-term goals should have been the first step in US Afghan policy.
Expedient vs. Realistic Timelines
It is essential to consider the time required for policy success. This is particularly true when establishing a new form of government or rebuilding a society after a civil war. Unfortunately, this has not been the case with US policy, which tends to be driven by “politically feasible” timelines rather than ones designed to solve the problems at hand. The gap between these timelines needs to be examined and policy modified accordingly. Timelines considered politically inexpedient should not be rejected outright.
There are examples of successful change from corrupt autocracies into functioning democracies with strong armies. South Korea is an example of a country that moved from a corrupt, kleptocratic government to the democratic, economically successful one it is today. Taiwan is another such example. These cases suggest that decades are necessary for such change. While the United States spent 20 years in Afghanistan, it never had policy thresholds that extended beyond one administration. The result was, as John Paul Vann famously said in Vietnam, “We don’t have 12 years’ experience. We have one year’s experience 12 times.” In Afghanistan’s case, the US presence had one year’s experience 20 times.
A realistic understanding of the time needed for social change, anti-corruption measures, and democracy to take root could have led to commitments over a much longer period, perhaps with expenditures more drawn out and less concentrated in a year or two. Such a policy would have required very different public policies to explain the timelines and to build appropriate expectations for the pace of progress. Instead, the constant demand for rapid progress and the pretense that it was happening had the result of undercutting policy support over time.
Alternatively, if such commitment was not possible, a realistic understanding of essential timelines might have led to a variety of different ways to leave earlier, even if what we left behind was unsatisfactory. In any event, the refusal to look realistically at the time requirement meant that we were trapped in unrealistic policies over and over. This lesson is worth learning because a realistic appraisal of the time required for operational success will be necessary in the future.
The Importance of Local Leadership
In Afghanistan, as in Iraq and Vietnam, the United States found itself with local partners who were not up to the requirements of the situation. They could neither address the seeping corruption nor control infighting among their supporters. Consequently, the field was wide open to insurgents. Two potential lessons can be drawn from this problem.
One is that we will need to be realistic in judging whether we have local partners who are up to the broad requirements of whatever policy we are engaged in. They may exist—President Ramon Magsaysay in the Philippines was such a leader, and the result was a successful counterinsurgency and nation-building. In the absence of such leadership, the United States has a long record of trying to compensate by either making policy in Washington or deposing the leader. We attempted both in Vietnam and Afghanistan but failed in both places.
Many examples of the problem with local partners not meeting the needs of the situation are found in the book Vietnam Declassified: The CIA and Counterinsurgency. Over and over again, American officials identified problems and devised policies to address them but were unable to obtain lasting local buy-in. Whenever American officials rotated, funding ran out, or a particular local partner was killed or transferred, the situation returned to square one.
The same pattern repeated itself in Iraq and Afghanistan. The Obama surge in Afghanistan produced not only Provincial Reconstruction Teams (PRTs) but also District Support Teams (DSTs) and a significant effort to increase the number of civilian advisors. This was intended to produce a significant change in governance within a very short time. As observers noted at the time, the Afghan government lacked the institutional capacity to capitalize on the progress achieved.
The underlying problem was not only the lack of time but also the willingness of the political leadership in Afghanistan or Vietnam to make necessary changes on their own. The problem was aptly captured in the famous, leaked “NODIS” telegram from then-US Ambassador to Afghanistan Karl Eikenberry, who observed that the basic problem with the proposed strategy was that we had no local partner.
Without adequate partners on the ground, policy frequently fails. Equally consistent has been our reaction to the lack of local partners. Either we try to build our own policies, as described above, or we change the leadership. In South Vietnam, the United States supported a coup that resulted in President Ngo Dinh Diem’s murder in 1963. In Afghanistan, Ambassador Richard Holbrooke tried to remove President Hamid Karzai, an effort that failed and further alienated the Afghan president. The US approach is not only arrogant and mechanistic but also completely unsuccessful. After 70 years of consistent failure, it is time for policymakers and academics to understand that foreign policy cannot be made without regard to foreigners.
There will not be a single solution to this problem in the future. It may not even be clear that the problem exists until the United States is deeply immersed in a country and its choices are limited. However, addressing the issue will require acknowledging that the problem exists and debating solutions in both academic and policy forums. Discussing policy failure without examining the underlying attitudes and approaches of local leaders has been a repeated phenomenon in many different administrations and countries, yielding the same poor results. There are underlying problems like these that extend beyond individual policy choices. Recognizing the importance of local leadership would be a starting point for making better choices in the future.
Policing: Paramilitary or Civil?
Building an effective police force has been a key issue in the insurgencies America has confronted. Professor Howard’s book also noted that police training is among the most challenging problems across various UN missions. When a problem recurs repeatedly, it is time to consider whether there is a deeper issue beyond the operational or organizational decisions in a particular country.
Several key points must be understood to develop a new approach to police training. One is that the United States is particularly badly placed for police training. We have no national police force. We have no national doctrine for police training. We have no established source of recruitment for police trainers, except for a limited number of retired police officers. Most active police forces do not want to give up their personnel to foreign missions.
In Afghanistan and Iraq, the debilities were exacerbated by the argument over whether the police force should be more along paramilitary or civilian lines. Of course, the answer was that both were necessary. Without proper law enforcement training, the police force could not act as a source of justice or public safety. However, the Afghan police still had to face large, heavily armed insurgent groups.
Mixed civil and paramilitary forces do exist in France, Italy, and Spain, but not in the United States. But these examples never made it into police training in Afghanistan. The international training mission drew Italy and France into training the Afghan border police, but not for regular police training.
The time given for police training was also too brief. In the United States, the average time for police training is 21 weeks, and this training is typically provided to at least high school graduates in established police forces who are not involved in counterinsurgency operations. In Afghanistan, training rarely exceeded several months, with recruits who were frequently illiterate and unable to perform basic reporting or record-keeping tasks.
There are some examples of comparatively successful police training. Robert Perito’s book The American Experience with Police in Peace Operations notes the necessity of an adequate ratio of police trainers to police recruits. Nothing in the length of time US and international forces devoted to police training in Afghanistan (or Iraq, for that matter) suggests that we learned this lesson.
The need to strengthen a local police force may arise in many cases that do not involve counterinsurgency or state-building efforts. Hence, Afghanistan’s lessons still matter. The resort to an outmatched Kenyan police force in Haiti, a country overrun by armed gangs, does not suggest we have even tried to take this history to heart.
No doubt there are other important lessons to be learned from Afghanistan. To profit from them, academics, as well as policymakers, need to go beyond catchphrases. They will need to consider that repeated failures reflect deeper, structural problems in our approach. If we cannot solve the problems of local leadership quality or the need for realistic time horizons, we must at least begin by acknowledging that these problems exist and recur. Only then will we be able to formulate better approaches for the future.
Ronald E Neumann was the US Ambassador to Afghanistan from 2005 to 2007, as well as Ambassador to Algeria and Bahrain. He served as an infantry officer in Vietnam and a senior officer in Iraq (2004–2005).
Afghanistan Peace Campaign
