A Roadmap for Better Choices from Security Partners


How can America best work with imperfect partners to achieve pressing security objectives? This question bedevils American foreign policy. Case in point: Over the past decade, America’s most essential military partner in Africa has become one of its most problematic. Uganda has played a key role in two major U.S. priorities on the continent — the African Union Mission in Somalia and the multinational campaign against Joseph Kony’s Lord’s Resistance Army. Even while Uganda’s involvement in these initiatives has led the United States to make it one of the largest recipients of American aid on the continent, its military has been accused of corruption, political interference, and serious human rights violations. The United States has sought to use diplomatic tools to redirect Ugandan behavior, including suspending some aid and canceling a military exercise in 2014, with little evidence of success.

Security sector assistance is the range of activities — including training, advising, equipping, exercising, hosting educational exchanges, and institutional capacity building — used by the U.S. government to enable foreign partners to effectively, transparently, and accountably govern their own security sectors. The United States implements security sector assistance primarily through the Departments of State and Defense, though the Departments of Justice, Homeland Security, the Treasury, and others contribute programs, personnel, and resources as well. The State Department provides security sector assistance to foreign militaries, law enforcement organizations, non-military security agencies such as border patrols and port security, and judicial systems; the Pentagon primarily provides assistance to foreign militaries, though it has authorities to assist non-military security forces in some cases.



Clear data on the scope of security sector assistance is difficult to come by. The Congressional Research Service estimates that the State Department spends roughly $7.5 billion annually on security assistance, the portion of security sector assistance directed to foreign militaries (though this estimate excludes non-military security sector assistance), and that the Pentagon executes about $9.5 billion each year (though Overseas Contingency Operations funds for Afghanistan and Iraq make up a substantial majority of this number, while several Pentagon programs avoid reporting expenditures). Accounting challenges aside, security sector assistance has been relied upon around the world to achieve a wide range of security objectives, especially since the Sept. 11, 2001, terrorist attacks: From drawing down and ending U.S.-led campaigns (e.g., Iraq and Afghanistan) to tilting the battlefield in regional conflicts (e.g., Ukraine and Syria) to deterring future conflicts (e.g., South China Sea and Northern and Eastern Europe), security sector assistance has taken center stage.

But it hasn’t worked out as Washington hoped. Instead of reducing risks and bringing America closer to its goals, we’ve seen suboptimal results, undermining U.S. strategies and heightening risk to U.S. interests.

As policymakers continually search for new approaches to increase the return on U.S. investments in security sector assistance, attention often turns to the idea of political conditionality. Political conditionality is generally understood as the application of incentives and disincentives tied to identifiable actions by a recipient of assistance to influence the behavior of that recipient. Thus far, the United States has not developed a framework for how to apply conditions systematically to achieve concrete outcomes. America’s historical experience with conditionality indicates that the use of positive conditions, in particular, could improve results in shaping partner behavior. A concrete model for applying positive conditionality to undergird security sector assistance and incentivize desired partner actions can help planners and policymakers realize the potential of security sector assistance to achieve strategic objectives.

A History of Conditions

The United States has adopted forms of political conditionality since at least 1974, when the U.S. Congress passed legislation stating that “no security assistance may be provided to any country the government of which engages in a consistent pattern of gross violations of internationally recognized human rights.” Since then, the United States has adopted conditionality in a number of cases, including banning security sector assistance to Indonesia from 1992 through 2005 due to human rights concerns, conditioning security sector assistance to Egypt based on concerns about human rights and democratic governance beginning in 2017, and limiting assistance to Pakistan from 1990 to 2001 and again beginning in 2017 due to concerns about its nuclear proliferation and support for terrorism.

Conditionality is a rich and complex concept, without consensus definitions among scholars. Svea Koch argues that “conditionality thus describes an incentive instrument in the relationship between two actors, in which one actor aims at changing the behavior of the other by setting up conditions for the relationship and by manipulating its cost-benefit calculation by using material incentives.” She proposes a useful four-part typology, in which conditionality can be characterized by whether it is ex ante (that is, whether conditions need to be met before an assistance relationship is initiated) or ex post (in which conditions are agreed or imposed during the course of an existing assistance relationship), as well as by whether it is positive (“rewarding and incentivizing”) or negative (“punitive and restrictive”).

In the realm of security sector assistance, experience has been disproportionately focused on negative approaches, almost always negative and ex post. In addition to Egypt, Indonesia, and Pakistan, the United States has sought to limit or cut off security sector assistance in El Salvador in the 1980s, Rwanda in 2012, Bahrain in 2011, and Burundi in 2015. All these efforts followed the ex post/negative pattern in which the United States has taken essentially punitive measures to reframe an existing security sector assistance relationship. One might view the application of the Leahy Law, which restricts assistance in cases of human rights abuse, or post-coup legal restrictions as cases of ex ante/negative conditionality, given that partners are expected to be aware of these condition-triggering laws prior to the acceptance of assistance or the commission of violations.

By contrast, examples of positive conditionality in the security sector assistance arena have been few and far between, even though the idea is not particularly new. “Incentive-based conditions” might tie additional and/or more sophisticated levels of security sector assistance, including new military capabilities or specific weapons platforms, to reform milestones to incentivize behavioral changes. In addition, rather than authorizing aid and then withholding it, on the basis of shared interests, the United States might identify positive actions that a recipient is considering taking and then incentivize them. More recently, the U.S. Institute for Peace Task Force on Extremism in Fragile States argued for “graduated security sector assistance,” in which the United States calibrates its engagement with fragile states according to a jointly agreed plan that enables partners in those countries to achieve escalating levels and types of U.S. security sector assistance as partners achieve reform milestones leading to more inclusive governance. These recommendations build on recent experiments in new government approaches to security sector assistance, such as the Security Governance Initiative, which have sought to establish joint bilateral planning on the basis of overlapping interests as the foundation of assistance programs, though these initiatives have not ultimately meaningfully conditioned assistance. The challenge has been that, while positive conditionality has been posited as a potentially viable and beneficial approach to improving security sector assistance outcomes, it has neither been sufficiently attempted in practice nor sufficiently elaborated in literature.

Toward a Framework for Positive Conditionality

What if the United States had approached its relationship with Uganda differently in pursuit of shared objectives? It’s possible that if the United States had structured its assistance in a way that incentivized Uganda’s politicians and security officials to take specific steps toward clearly — and mutually — identified milestones, that could have led to the development of stronger capabilities for and commitment to civilian protection and governance by Uganda’s military over time. Instead, a reactive posture to human rights and corruption transgressions has had limited effects, and punitive measures may only strain the relationship and undermine possibilities for cooperation.

The United States should develop a framework for positive conditionality in security sector assistance to better shape political and security outcomes with partner countries. On balance an affirmative framework will offer the United States the widest latitude in shaping outcomes in a broad range of circumstances. This approach also does not rule out the sequencing of punitive steps — or a pause to assess partner performance.

First, this framework assumes an ex ante/positive approach to conditionality, at least in some manner. With the U.S. government providing security sector assistance to nearly 200 countries around the world, it is unlikely that the United States will be initiating a security sector assistance relationship wholly from scratch. An ex ante approach would focus on developing concrete plans, including triggers for clearly identified conditions, at the inception of an initiative or set of programs. It should involve a plan covering at least five years, identifying ultimate objectives and intermediate milestones, a theory of change for how objectives will be achieved, and metrics to provide a basis for assessing progress. Policymakers and planners can then link conditions to milestones and metrics. Moreover, such an approach builds on recognized best practices for capacity-building initiatives.

Second, the framework must include a partner government’s mutual participation in the creation of the assistance plan, the identification of objectives, and the agreement on conditions. The partner will be more incentivized to progress toward objectives when it understands and commits to such objectives based on its own identified interests. A memorandum of understanding, bilateral compact, or some other formal written instrument can commit both parties to its terms.

Third, the framework should structure conditions as positive inducements for the recipient to take steps toward milestones or objectives. Milestones could include completion of defense institutional reforms, starting with development of a process to align budget to strategy, progressing to completion of a first budget cycle with the new process, and culminating in institutionalization of the process through policy and/or law. They might also include progress toward capability or interoperability milestones (e.g., ability to conduct nighttime counter-terrorism operations in accordance with the laws of armed conflict or command and control and intelligence integration across platforms and systems). In addition, milestones could include transparency and accountability activities (e.g., publication of public budget, completion of audit, investigation of specific incidents of resource waste and abuse or diversion, establishment and activity of an independent third-party investigative unit for accountability).

Partners will likely respond to inducements that are material in nature, including access to an expanded variety of types of assistance and an expanded quantity of assistance. However, partners may also value incentives such as eligibility for key agreements to facilitate access and information and/or technology sharing (e.g., acquisition and cross-servicing, general security of military information, etc.) or eligibility for expanded partnership opportunities (e.g., National Guard State Partnership Program or the Military Personnel Exchange Program). Moreover, incentives may also be political, such as access to membership in certain organizations or eligibility for certain agreements.

Such inducements are consistent with the Task Force on Extremism in Fragile States’ concept of graduated assistance: As a partner achieves key milestones, it is able to “graduate” into a new status that allows it to access increasing quantities and types of assistance, as well as other benefits. Policymakers should tailor the potential tiers of access and associated benefits, as well as the milestones triggering graduation to a higher tier, to the individual recipient, and accordingly integrate them into bilateral development plans or compacts, as discussed above. A possible tiering of graduated assistance could include:

  • Tier One: nonlethal training and individual education.
  • Tier Two: comprehensive training and nonlethal equipment.
  • Tier Three: lethal equipment and access to U.S. military staff colleges.
  • Tier Four: high-level exercises, high-level staff talks, and joint and combined planning and operations.

Yet, for fragile states with weak security institutions, the task force rightly notes that assistance should initially focus on institutional reforms, with “increasingly sophisticated and lethal capabilities and equipment … contingent on sustained improvements in security sector governance.” The reforms should also institutionalize inclusivity. In practice, such an approach may prove difficult if partner governments or their security forces are vested in a corrupt status quo, such that institutional reform poses an existential threat or compromise to their personal gains. In such partnerships, identifying geographic areas or bureaucratic processes that may be more receptive to change first should be the focus of investment, and may be most successful if there are third parties, such as other allies and partners, media, or civil society that are calling for similar changes. For a potential coalition partner with comparatively good governance, on the other hand, milestones and incentives may be built around coalition interoperability and the capability to perform certain roles within the coalition context. In collaboration with the partner, U.S. policymakers should design a program of graduated or tiered assistance, structuring incentives and milestones across the full portfolio of security sector assistance. The total program of assistance could grow in both scope and sophistication as the partner progresses through milestones.

In addition, within each broad category of security sector assistance provided by the United States, policymakers and planners may find opportunities to moderate U.S. investments based on a partner’s progress (in other words, in relation to specific security sector assistance lines of effort). The table below outlines engagement at low, moderate, and high levels across several broad assistance categories as they could be applied within the U.S. Indo-Pacific Command area of responsibility. As planners construct graduated assistance programs, they thus have options to moderate assistance both within a specific type of support and across the full portfolio of security sector assistance. In practice, the categories and degrees of security sector assistance options would be tailored to meet the objectives of the specific partner relationship.

Figure 1: Potential Categories and Degrees of Security Sector Assistance via Indo-Pacific Command

  • Visit by U.S. Army Pacific Commander or J-4
  • Visit by Commander or Service Secretaries
  • Visit by the Secretary of Defense or the President
  • Unit-level (e.g., Submarine Group 7) staff talks
  • Port visits
  • Attendance at conferences and symposiums
  • Staff talks with U.S. 7th Fleet
  • Personnel exchanges
  • Assignment of liaison officer
  • Staff talks with U.S. Pacific Fleet
  • Joint patrols


  • Observation of Rim of the Pacific Exercise
  • Participation in Cobra Gold
  • Hosting Cooperation Afloat Readiness and Training exercises
  • Observation of U.S. Exercises (e.g., Red Flag)
  • Hosting significant bilateral exercise
Institutional Capacity-Building
  • Inouye Center Seminars on Defense Governance
  • Civilian personnel exchanges
  • Periodic Defense Institution Reform Initiative seminars on defense governance
  • Ministry of Defense Advisor assignment
  • Long-term, intensive Defense Institution Reform Initiative program
Materiel Assistance
  • Unit training
  • Smaller volume, less sophisticated materiel assistance
  • Nonlethal assistance


  • Lethal assistance
  • Higher volume, more sophisticated materiel assistance
  • MOU on research collaboration
  • MOU on planned assistance
  • Acquisition and Cross-Servicing Agreement
  • General Security of Military Information Agreement
  • Status of Forces Agreement
  • Defense cooperation agreement

Importantly, conditions should be based on realistic assumptions about what partners would be willing to do (in terms of both what steps can be realistically expected given the recipient’s unique political and institutional context as well as what leverage relatively small sums of U.S. assistance can be expected to produce). Also, conditions should be accompanied by concrete U.S. commitments. Positive conditionality will quickly cease to be attractive to partners when the United States fails to follow through on a promised inducement.

Of course, following through in this manner requires sustained availability of funding. Many security sector assistance programs have significant flexibility in the availability of appropriated funds over time. More importantly, in our experience, programs with more rigor in strategy and long-term planning are likely to win the sustained support of congressional appropriators.

This proposed model of positive conditionality offers new possibilities not only for incenting partners to take steps toward shared objectives but also for improving planning and management of security sector assistance resources. Assistance programs are structured such that the achievement of each milestone triggers an expansion of the assistance relationship. Each step also provides an opportunity for an assessment of goals and progress. It enables planners to incorporate natural mechanisms to provide exit ramps, through which the U.S. or partner government can determine to terminate the initiative, or to restructure efforts to achieve a defined milestone. To be effective, it is critical that overarching objectives, as well as milestones leading toward those objectives, are specific, measurable, and time bound.

Deciding When to Use Positive Conditionality

Given competing political pressures and security obligations, positive conditionality will not be the best approach for every security sector assistance relationship. So, when should it be applied? Until implementers of security sector assistance can experiment with positive conditionality in diverse contexts in the field, it will be difficult to provide detailed answers, but we can offer a few hypotheses. The first is that the effectiveness of positive conditionality in the security sector will depend on the degree of recipient country ownership. According to the World Bank, country ownership “means that there is sufficient political support within a country to implement its developmental strategy, including the projects, programs, and policies for which external partners provide assistance.” Recipients will be more likely to take desired actions to the extent they share an initiative’s objectives. Perhaps more importantly, though, as research in the development community has demonstrated, conditionality can undermine the effectiveness and legitimacy of governing institutions if conditions are imposed in ways that supplant domestic decision-making processes or force domestic actions without the recipient government’s buy-in. To be effective, positive conditionality should reinforce domestic political institutions, not undermine them.

A second hypothesis is that positive conditionality will be most effective when structured around a recipient’s concrete priority. In many cases, partners have goals and aspirations that the United States can substantially influence; for example, accession to or interoperability with NATO, participation in military or peacekeeping coalitions, access to certain sophisticated military systems, or certain political agreements. To be effective, positive conditionality should structure milestones and incentives around such concrete priorities determined by the partner, rather than ambiguously defined partnership or capability goals or objectives imposed by the United States. Positive conditionality will work best when the United States is uniquely positioned, in relation to other security exporters, to help a partner obtain its goals.

Third, positive conditionality will work best when the donor avoids interest asymmetry. If positive conditionality is to succeed, the United States cannot afford to need a security relationship more than the recipient. In Uganda, as well as in some other notable cases in recent years, the United States has become so reliant upon single partners (e.g., the security forces of Iraq and Afghanistan; Saudi Arabia in Yemen) that its policy options were sharply constrained by the negative consequences of terminating or redirecting assistance, meaning that Washington was unable to enforce the consequences of noncompliance. Such conditions are incompatible with the model we propose.

Finally, we should acknowledge that any attempt to implement positive conditionality will not take place on a green field; it should be crafted within the context of existing U.S. assistance and arrangements that will limit options and complicate negotiations. In addition, the policy and resourcing process within the executive branch, tensions and different priorities between and among regional and functional offices, as well as interactions and priorities of Congress will add further complexity to implementing a conditionality framework. Simply put, the United States has already committed to assistance and agreements around the world without seeking concomitant commitments from partners. Elements of the U.S. government have different viewpoints on how assistance should be leveraged and applied. This legacy does not foreclose opportunities for positive conditionality but does require a creative and flexible application of the model. Agreements in place can be renegotiated; existing assistance can be phased out over time to make way for new priorities and new approaches. Baseline assessments can help identify a partner’s most pressing needs and those priorities that may be most appealing for future work.

Some will object that this approach and the obligations it implies for both U.S. planners and foreign counterparts are unrealistic, and that such obligations attached to U.S. assistance could drive partners to alternative security providers. In our experience, many U.S. partners not only want to work with the United States as their security partner of choice but are also motivated by a desire to improve the effectiveness and governability of their own security sectors. Yet, it is also important to acknowledge that this approach will unquestionably require more diplomatic engagement by senior policymakers to achieve and sustain partner buy-in. That is both necessary and appropriate. Security sector assistance is a fundamentally political activity. It should be buttressed by engagement between political stakeholders on both sides to identify political objectives, requirements, and obstacles in relation to an assistance program. Indeed, the lack of this type of persistent political dialogue is a prominent cause of many security sector assistance failures.

Looking Ahead

Incentivizing security partners’ progress toward common outcomes and objectives will enable the United States to get more out of its security sector assistance investments. Such an approach will be vital to building strong, accountable, and interoperable partnerships globally. It may be particularly important in fragile states, where security sector assistance may otherwise inadvertently fuel predatory behavior by security actors. While political realities will inevitably disrupt routine plans, and policymakers will demand instant or tangible “deliverables,” such as lethal assistance, the only remedy to mitigate against unintended consequences is to institutionalize a smart, incentives-based planning framework with continuous opportunities for adaptation and adjustment. The United States will need to pilot this approach to experiment and fine-tune a viable framework. This in turn will require bipartisan support in the U.S. Congress for the security sector assistance enterprise to “fail forward” on a pathway to success.

Several indicators suggest that Congress is ready to move in this direction. In recent years, it has passed several bills — most notably the Fiscal Year 2017 National Defense Authorization Act and the Foreign Aid Transparency and Accountability Act of 2016 — that have established important foundations for positive conditionality, including mandates for assessment, monitoring, and evaluation of foreign military assistance, more flexible planning horizons for Defense Department assistance, and tighter collaboration between the Defense and State Departments in planning and implementing military aid.

A new legislative initiative, the pending Department of State Authorization Act of 2019, would take several additional steps forward. Recently passed by the House Foreign Affairs Committee with strong bipartisan support, the bill would create several new tools supporting positive conditionality, including strengthening assessment, monitoring, and evaluation of military assistance, authorizing “Security Assistance Compacts” similar to the bilateral compacts discussed above, and bolstering planning processes to achieve greater strategic focus and prioritization in U.S. assistance. Passage of this legislation would represent a watershed moment for U.S. security sector assistance, particularly aid provided by the Department of State.

The urgency of adopting a new, more strategic approach to security assistance is growing by the day. American partnerships are under increasing pressure from declining resources, domestic isolationist trends, and strategic competition. Meanwhile, the range of global security challenges for which the United States turns to partners to contribute grows in diversity and complexity. Positive conditionality offers policymakers, diplomats, and operators a tool to shape partners’ behaviors in a manner that offers an opportunity to produce greater return on American investments, simultaneously strengthening partnerships and growing global capacity to deter, prevent, and respond to crises and conflict.



Tommy Ross is a non-resident senior associate at the Center for Strategic and International Studies. He served as Deputy Assistant Secretary of Defense for Security Cooperation at the Pentagon and was the senior defense and intelligence adviser to Senate Majority Leader Harry Reid. He has also held other senior positions in the House and Senate.

Melissa Dalton directs the Cooperative Defense Project and is a senior fellow and deputy director of the International Security Program at the Center for Strategic and International Studies. Prior to CSIS, she served in the U.S. Department of Defense for 10 years.

Image: U.S. Marine Corps (Photo by Staff Sgt. Mark E Morrow Jr.)